Retracted papers still cited on English Wikipedia

The model's reasoning: The evidence assessment draws on the review's finding that only one trial supplied trazodone data, which were excluded from its meta-analysis; the other reviews plausibly support the broader assessment independently.

The sentence the footnote ends: However, use of trazodone in anxiety disorders is off-label and evidence of its effectiveness for these indications is variable and limited.

Other citations on it:

• Bossini L, Coluccia A, Casolaro I, Benbow J, Amodeo G, De Giorgi R, et al. (2015). "Off-Label Trazodone Prescription: Evidence, Benefits and Risks". Curr Pharm Des. 21 (23): 3343–51. doi:10.2174/1381612821666150619092236. PMID 26088119.
• Bossini L, Casolaro I, Koukouna D, Cecchini F, Fagiolini A (August 2012). "Off-label uses of trazodone: a review". Expert Opin Pharmacother. 13 (12): 1707–17. doi:10.1517/14656566.2012.699523. PMID 22712761. S2CID 10638259.
• Khouzam HR (2017). "A review of trazodone use in psychiatric and medical conditions". Postgrad Med. 129 (1): 140–48. doi:10.1080/00325481.2017.1249265. PMID 27744763. S2CID 205452630.
• Asnis GM, Kohn SR, Henderson M, Brown NL (2004). "SSRIs versus non-SSRIs in post-traumatic stress disorder: an update with recommendations". Drugs. 64 (4): 383–404. doi:10.2165/00003495-200464040-00004. PMID 14969574. S2CID 38329753.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.86. Finding RW-31637 / 2326240 / #1, read from revision 1377081496 of Trazodone.

The model's reasoning: This states the paper's reported graft preparation as fact, rather than its falsified success claims; the newspaper plausibly supports the procedural history, but the follow-up paper is also retracted.

Where the model says the article warns about this paper:

• Later (Retractions): The following papers authored by Macchiarini have been retracted:
• Later (Retractions): October 2023, 2008 Lancet paper, retracted by the journal for falsification, including the claim that 'the graft immediately provided the recipient with a functional airway, improved her quality of life, and had a normal appearance and mechanical properties at 4 months':

Other citations on it:

• Cowell, Alan; Grady, Denise (19 November 2008). "Europeans Announce Pioneering Surgery". The New York Times.
• Gonfiotti, Alessandro; Jaus, Massimo O; Barale, Daniel; Baiguera, Silvia; Comin, Camilla; Lavorini, Federico; Fontana, Giovanni; Sibila, Oriol; Rombolà, Giovanni; Jungebluth, Philipp; Macchiarini, Paolo (January 2014). "Retracted: The first tissue-engineered airway transplantation: 5-year follow-up results". The Lancet. 383 (9913): 238–244. doi:10.1016/S0140-6736(13)62033-4. PMID 24161821. S2CID 41597285. (Retracted, see doi:10.1016/S0140-6736(23)02340-1, PMID 37898524, Retraction Watch)

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.94. Finding RW-48567 / 24365639 / #1, read from revision 1375891403 of Paolo Macchiarini.

The model's reasoning: The quantified reduction closely matches the review's 36% mortality reduction in patients with coronary heart disease, but the article presents it as a general risk reduction; the NHS prevention page may support cessation advice, but support for this percentage is unclear.

The whole list item: Stopping smoking and avoidance of second-hand smoke. Stopping smoking reduces risk by about 35%.

Other citations on it:

• "Heart Attack—Prevention". NHS Direct. 28 November 2019.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-33394 / 512662 / #1, read from revision 1377545070 of Cardiovascular disease.

The model's reasoning: This benefit–harm assessment states the review's conclusion as current evidence; the Choosing Wisely recommendation plausibly supports the same assessment, and the review was retracted and replaced.

The sentence the footnote ends: Most sinusitis infections are caused by viruses, for which amoxicillin and amoxicillin-clavulanate are ineffective, and the small benefit gained by amoxicillin may be overridden by the adverse effects.

Other citations on it:

• "Five Things Physicians and Patients Should Question" (PDF). Choosing Wisely: An Initiative of the ABIM Foundation. American Academy of Allergy, Asthma, and Immunology. Archived (PDF) from the original on 3 November 2012. Retrieved 14 August 2012.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-31898 / 2885 / #1, read from revision 1363961139 of Amoxicillin.

The model's reasoning: The attributed US mortality estimate is presented without a retraction warning and is supported only by Uva; the supplied retraction reason is incomplete, so its effect on the estimate requires source review.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-4358 / 1195328 / #1, read from revision 1369818762 of Erotic asphyxiation.

The model's reasoning: The paper reports TNF-induced caspase-independent apoptosis through a LAPF-dependent lysosomal–mitochondrial pathway; this sentence presents that conclusion as a biological possibility without noting the retraction.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-29253 / 2457 / #1, read from revision 1373653552 of Apoptosis.

The model's reasoning: This reports the retracted review's conclusion as valid evidence; the other sources address therapeutic touch generally and do not clearly support the specific acute-wound conclusion.

The whole list item: Therapeutic touch – is a form of vitalism where a practitioner, who may be also a nurse, passes their hands over and around a patient to "realign" or "rebalance" a putative energy field. A recent review concluded that "[t]here is no evidence that [Therapeutic Touch] promotes healing of acute wounds." No biophysical basis for such an energy field has been found.

Other citations on it:

• ""Beyond Science", on season 8, episode 2". Scientific American Frontiers. Chedd-Angier Production Company. 1997–1998. PBS. Archived from the original on 1 January 2006.
• Wallace, Sampson; Vaughn, Lewis (24 March 1998). "'Therapeutic Touch' Fails a Rare Scientific Test". CSICOP News. Committee for Skeptical Inquiry. Archived from the original on 13 October 2007. Retrieved 5 December 2007. Despite this lack of evidence, TT is now supported by major nursing organizations such as the National League of Nurses and the American Nurses Association.
• "Scientific American". Archived from the original on 9 October 2009.
• Courcey, Kevin. "Further Notes on Therapeutic Touch". Quackwatch. Retrieved 5 December 2007. What's missing from all of this, of course, is any statement by Krieger and her disciples about how the existence of their energy field can be demonstrated by scientifically accepted methods.
• "Energy Medicine: An Overview". National Center for Complementary and Integrative Health. 24 October 2007. Retrieved 5 December 2007. neither the external energy fields nor their therapeutic effects have been demonstrated convincingly by any biophysical means.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-31160 / 267014 / #1, read from revision 1377114396 of List of pseudoscience topics.

The model's reasoning: The sole citation is the retracted study; this treatment-response claim also goes beyond its abstract, which reports an association between prior COVID-19 and DSPD rather than impaired recovery or treatment efficacy.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.95. Finding RW-68660 / 231014 / #1, read from revision 1378143242 of Delayed sleep phase disorder.

The model's reasoning: Listing long COVID as a risk factor appears to summarize the retracted study's COVID-19–DSPD association; the other listed risk factors are unrelated, and an editor should check whether long COVID has independent sourcing.

The model's confidence: 0.85. Finding RW-68660 / 231014 / #2, read from revision 1378143242 of Delayed sleep phase disorder.

The model's reasoning: This repeats the COVID-19–DSPD relationship reported by the retracted study cited in the next paragraph; an editor should check whether this paragraph has an independent source.

Where the model found it: COVID-19, caused by the SARS-CoV-2 virus, has been observed to act as a comorbidity or aggravating factor in DSPD. Post-acute sequelae of COVID-19 (commonly referred to as long COVID) often include neurological and autonomic dysregulation, which can interfere with circadian timing systems. These effects may involve delayed melatonin release, disrupted cortisol rhythms, and autonomic nervous system instability, all of which can exacerbate or prolong DSPD symptoms.

The model's confidence: 0.91. Finding RW-68660 / 231014 / #3, read from revision 1378143242 of Delayed sleep phase disorder.

The model's reasoning: This states the paper's survival and neurological-outcome findings for epinephrine versus placebo or no drugs; the accompanying updated meta-analysis plausibly supports the same claim.

Other citations on it:

• Aves T, Chopra A, Patel M, Lin S (27 November 2019). "Epinephrine for Out-of-Hospital Cardiac Arrest: An Updated Systematic Review and Meta-Analysis". Critical Care Medicine. 48 (2): 225–229. doi:10.1097/CCM.0000000000004130. PMID 31789700. S2CID 208537959.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-21798 / 60575 / #1, read from revision 1378042461 of Cardiac arrest.

The model's reasoning: The paragraph explicitly derives these conclusions from the study above and repeats its preventive-therapy claim without a local retraction warning; an editor should check whether its own sources independently support these conclusions.

Where the model found it: From the study above, several conclusions have been made. A relationship between personality and cancer exists, along with a relationship between personality and coronary heart disease. Personality type acts as a risk factor for diseases and interacts synergistically with other risk factors, such as smoking and heredity. It has been statistically proven that behavior therapy can significantly reduce the likelihood of cancer or coronary heart disease mortality. Studies suggest that both body and mental disease arise from each other. Mental disorders arise from physical causes, and likewise, physical disorders arise from mental causes. While Type A personality did not show a strong direct relationship between its attributes and the cause of coronary heart disease, other types of personalities have shown strong influences on both cancer-prone patients and those prone to coronary heart disease.

Where the model says the article warns about this paper:

• In the same section (Other studies): studies (that also later were questioned for non plausible results were considered unsafe, and were retracted in 2021)

The model's confidence: 0.96. Finding RW-29959 / 1699268 / #2, read from revision 1376804284 of Type A and Type B personality theory.

The model's reasoning: This repeats the paper's central transatlantic public-opinion comparison; an editor should check whether this paragraph has an independent source.

Where the model found it: There are large differences in consumer acceptance of GMOs, with Europeans more likely to view GM food negatively than North Americans. GMOs arrived on the scene as the public confidence in food safety, attributed to recent food scares such as Bovine spongiform encephalopathy and other scandals involving government regulation of products in Europe, was low. This along with campaigns run by various non-governmental organizations (NGO) have been very successful in blocking or limiting the use of GM crops. NGOs like the Organic Consumers Association, the Union of Concerned Scientists, Greenpeace and other groups have said that risks have not been adequately identified and managed and that there are unanswered questions regarding the potential long-term impact on human health from food derived from GMOs. They propose mandatory labeling or a moratorium on such products.

The model's confidence: 0.88. Finding RW-889 / 12339 / #2, read from revision 1374510745 of Genetically modified organism.

The model's reasoning: The paragraph reports the paper’s arsenic-substitution finding without explicitly stating its retraction or refutation; attribution and “allegedly” do not supply that warning.

The sentence the footnote ends: This claim, published in Science, suggested that arsenic could potentially serve as a building block of life in place of phosphorus, challenging long-standing assumptions about biochemical requirements for life on Earth.

Where the model says the article warns about this paper:

• In the same section (Bacteria): Subsequent studies provided evidence contradicting the initial findings. One follow-up study published in Science in 2011 demonstrated that GFAJ-1 still requires phosphate to grow and does not incorporate arsenate into its DNA in any biologically significant way. Another independent investigation in 2012 used more sensitive techniques to purify and analyze the DNA of GFAJ-1 and found no detectable arsenate incorporated into the DNA backbone. The authors concluded that the original observations were likely due to experimental contamination or insufficient purification methods.
• In the same section (Bacteria): In 2025, the journal Science formally retracted the original paper, citing a lack of sufficient experimental support for its key conclusions, though the authors continued to stand by their data.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.98. Finding RW-66335 / 897 / #1, read from revision 1376412064 of Arsenic.

The model's reasoning: The caption presents regional economic damage projections from the paper as valid; the retraction concerns errors in the data and analyses and issues with the methods.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-72836 / 2119174 / #1, read from revision 1378412069 of Effects of climate change.

The model's reasoning: The nasal-irrigation benefit overlaps the review's conclusion and may summarize it in the lead; an editor should check whether this paragraph has independent support.

Where the model found it: Prevention of sinusitis focuses on regular hand washing, staying up-to-date on vaccinations, and avoiding smoking. Analgesics such as naproxen, nasal steroids, and nasal irrigation may be used to help with symptoms. Recommended initial treatment for acute sinusitis is watchful waiting. If symptoms do not improve in 7–10 days or worsen, then an antibiotic may be implemented or changed. In those in whom antibiotics are indicated, either amoxicillin or amoxicillin/clavulanate is recommended first line, with amoxicillin/clavulanate being superior to amoxicillin alone but with more side effects. Surgery may be recommended in those with chronic disease who have failed medical management.

The model's confidence: 0.72. Finding RW-31614 / 28598 / #2, read from revision 1373933225 of Sinusitis.

The model's reasoning: This table row repeats a symptomatic benefit of nasal irrigation consistent with the review; an editor should check its sources because independent support is not visible here.

The model's confidence: 0.70. Finding RW-31614 / 28598 / #3, read from revision 1373933225 of Sinusitis.

The model's reasoning: This closely repeats the paper's conclusion that a bowl provides about 60% of children's vitamin A requirement; an editor should check whether this paragraph has an independent source.

Where the model found it: Vitamin A deficiency is usually coupled to an unbalanced diet. Since carotenes are hydrophobic, sufficient fat must be present in the diet for golden rice (or most other vitamin A supplements) to alleviate vitamin A deficiency. Moreover, this claim referred to an early cultivar of golden rice; one bowl of the latest version provides 60% of Recommended Dietary Allowance (RDA) for healthy children. The RDA levels advocated in developed countries are far in excess of the amounts needed to prevent blindness.

Where the model says the article warns about this paper:

• Later (Research): However, in 2015, the journal retracted the study, claiming that the researchers had acted unethically when providing Chinese children golden rice without their parents' consent.

The model's confidence: 0.96. Finding RW-948 / 246458 / #4, read from revision 1374816022 of Golden rice.

The model's reasoning: This assessment reflects the review's conclusion that only weak evidence supported aciclovir for primary herpetic gingivostomatitis; the review was withdrawn as out of date and requires a current replacement source.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-31672 / 21323216 / #1, read from revision 1375958951 of Herpes.

The model's reasoning: This reports the withdrawn review's conclusions about insufficient evidence and serious risk of bias; withdrawal as out of date calls for an up-to-date source rather than removal on grounds of invalid findings.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-30910 / 28677895 / #1, read from revision 1376684169 of Transcendental Meditation technique.

The model's reasoning: The efficacy uncertainty reflects the review's conclusion that trials were insufficiently powered; the other therapeutic reviews plausibly support this claim, but the withdrawn, outdated review should be replaced.

The whole list item: Ribavirin is an antiviral medication licensed for the treatment of RSV in children. It is a guanosine analog that acts by inhibiting viral RNA synthesis and capping. It was approved in 1986 for treatment of RSV infection. However, the use of ribavirin remains controversial due to unclear evidence of efficacy and concerns about toxicity to exposed staff members, as well as cost. As such, treatment guidelines do not make recommendations for its use in children. In adults, ribavirin is used off-label and is generally reserved for the severely immunocompromised, such as those undergoing hematopoietic stem cell transplants.

Other citations on it:

• Simões EA, DeVincenzo JP, Boeckh M, Bont L, Crowe JE, Griffiths P, et al. (March 2015). "Challenges and Opportunities in Developing Respiratory Syncytial Virus Therapeutics". Journal of Infectious Diseases. 211 (suppl 1): S1–S20. doi:10.1093/infdis/jiu828. PMC 4345819. PMID 25713060.
• Jha A, Jarvis H, Fraser C, Openshaw PJ (June 2016). "Chapter 5: Respiratory Syncytial Virus". In Hui DS, Rossi GA, Johnston SL (eds.). SARS, MERS and other Viral Lung Infections. Wellcome Trust–Funded Monographs and Book Chapters. Sheffield (UK): European Respiratory Society. ISBN 978-1-84984-070-5. PMID 28742304. Archived from the original on 28 December 2020. Retrieved 29 October 2020.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.94. Finding RW-33569 / 477489 / #1, read from revision 1378374574 of Respiratory syncytial virus.

The model's reasoning: This reports the paper's periodic-pulse observation and neutron-star interpretation without noting that the findings were retracted.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-72832 / 40408388 / #1, read from revision 1378291915 of Fast radio burst.

The model's reasoning: Both sentences rely on the paper's claimed periodic components, duration and neutron-star interpretation, which are undermined by errors in data, methods and conclusions.

The sentence the footnote ends: A neutron star has been proposed as the origin of this FRB.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-72832 / 40408388 / #2, read from revision 1378291915 of Fast radio burst.

The model's reasoning: The paragraph presents the paper's event detection, measured pulse structure and unprecedented periodicity as valid findings without a retraction warning.

The sentence the footnote ends: This is the first time such a periodic pulse was detected.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-72832 / 40408388 / #3, read from revision 1378291915 of Fast radio burst.

The model's reasoning: The assertion that nutrition or supplementation is unlikely to help uses the overview as evidence about intervention efficacy, but its abstract reports insufficient robust evidence rather than established ineffectiveness; the removal reason supplies no further explanation.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.86. Finding RW-30943 / 400199 / #3, read from revision 1372897774 of Weight loss.

The model's reasoning: The paragraph presents the paper's DNMT1 recruitment, asymmetric methylation and approximately half-silenced recombinants as established mechanisms; the other sources plausibly support recruitment but need checking for the detailed methylation and expression claims.

The sentence the footnote ends: At least four articles report the recruitment of DNA methyltransferase 1 (DNMT1) to sites of DNA double-strand breaks.

Other citations on it:

• Mortusewicz O, Schermelleh L, Walter J, Cardoso MC, Leonhardt H (June 2005). "Recruitment of DNA methyltransferase I to DNA repair sites". Proceedings of the National Academy of Sciences. 102 (25): 8905–9. Bibcode:2005PNAS..102.8905M. doi:10.1073/pnas.0501034102. PMC 1157029. PMID 15956212.
• O'Hagan HM, Mohammad HP, Baylin SB (August 2008). "Double strand breaks can initiate gene silencing and SIRT1-dependent onset of DNA methylation in an exogenous promoter CpG island". PLOS Genetics. 4 (8) e1000155. doi:10.1371/journal.pgen.1000155. PMC 2491723. PMID 18704159.
• Ha K, Lee GE, Palii SS, Brown KD, Takeda Y, Liu K, et al. (January 2011). "Rapid and transient recruitment of DNMT1 to DNA double-strand breaks is mediated by its interaction with multiple components of the DNA damage response machinery". Human Molecular Genetics. 20 (1): 126–40. doi:10.1093/hmg/ddq451. PMC 3000680. PMID 20940144.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-69736 / 854294 / #1, read from revision 1376907584 of DNA repair.

The model's reasoning: This repair-associated demethylation claim is supported solely by the retracted paper; the abstract focuses on HR and does not establish the NHEJ-specific result, so the attribution requires source inspection.

The reference is tagged {{Retracted}}.

The model's confidence: 0.88. Finding RW-69736 / 854294 / #2, read from revision 1376907584 of DNA repair.

The model's reasoning: This states the withdrawn review's comparative-healing conclusion for venous ulcers; the other citation concerns diabetic foot ulcers, and the retract-and-replace notice calls for checking a replacement source.

The sentence the footnote ends: The results of meta-analyses indicate no significant difference in healing rates between hydrocolloid dressings and other dressings (including simple gauze dressings) for venous ulcers, or for diabetic foot ulcers.

Other citations on it:

• Dumville, Jo C.; Deshpande, Sohan; O'Meara, Susan; Speak, Katharine (6 August 2013). "Hydrocolloid dressings for healing diabetic foot ulcers". The Cochrane Database of Systematic Reviews. 2013 (8) CD009099. doi:10.1002/14651858.CD009099.pub3. ISSN 1469-493X. PMC 7111300. PMID 23922167.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-32404 / 5704220 / #2, read from revision 1368881437 of Hydrocolloid dressing.

The model's reasoning: This closely repeats the review's conclusion that carbamazepine has potential benefits but requires larger controlled studies; an editor should check whether the item has an independent, current source.

The model's confidence: 0.92. Finding RW-33038 / 11683572 / #2, read from revision 1373095815 of Benzodiazepine withdrawal syndrome.

The model's reasoning: The absence-of-evidence claim reflects the review's conclusion that no included studies assessed caries prevention; the review was withdrawn as out of date and replaced, rather than for invalid findings.

Other citations on it:

• de Oliveira K, Nemezio MA, Romualdo PC, da Silva R, de Paula E Silva F, Küchler EC (2017). "Dental Flossing and Proximal Caries in the Primary Dentition: A Systematic Review". Oral Health & Preventive Dentistry. 15 (5): 427–434. doi:10.3290/j.ohpd.a38780. PMID 28785751.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-30653 / 414350 / #1, read from revision 1369993076 of Tooth decay.

The model's reasoning: This repeats the review's conclusion of small to moderate symptom improvement with tamsulosin, generalized to alpha blockers; the other reviews plausibly support the claim, but this review was withdrawn as out of date.

Other citations on it:

• Hwang EC, Gandhi S, Jung JH, et al. (October 2018). "Naftopidil for the treatment of lower urinary tract symptoms compatible with benign prostatic hyperplasia". The Cochrane Database of Systematic Reviews. 2018 (10) CD007360. doi:10.1002/14651858.CD007360.pub3. PMC 6516835. PMID 30306544.
• Djavan B, Marberger M (1999). "A meta-analysis on the efficacy and tolerability of alpha1-adrenoceptor antagonists in patients with lower urinary tract symptoms suggestive of benign prostatic obstruction". European Urology. 36 (1): 1–13. doi:10.1159/000019919. PMID 10364649. S2CID 73366414.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-33406 / 88164 / #1, read from revision 1377554934 of Benign prostatic hyperplasia.

The model's reasoning: This reflects the review's findings of comparable efficacy and differing adverse effects among alpha antagonists; the other cited reviews plausibly support it, but this review was withdrawn as out of date.

Other citations on it:

• Hwang EC, Gandhi S, Jung JH, et al. (October 2018). "Naftopidil for the treatment of lower urinary tract symptoms compatible with benign prostatic hyperplasia". The Cochrane Database of Systematic Reviews. 2018 (10) CD007360. doi:10.1002/14651858.CD007360.pub3. PMC 6516835. PMID 30306544.
• Djavan B, Marberger M (1999). "A meta-analysis on the efficacy and tolerability of alpha1-adrenoceptor antagonists in patients with lower urinary tract symptoms suggestive of benign prostatic obstruction". European Urology. 36 (1): 1–13. doi:10.1159/000019919. PMID 10364649. S2CID 73366414.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-33406 / 88164 / #2, read from revision 1377554934 of Benign prostatic hyperplasia.

The model's reasoning: This reports the study's efficacy and harm conclusions without a retraction warning in this paragraph; the study is notable, but its data were unreliable.

The sentence the footnote ends: In May 2020, The Lancet published an observational retrospective cohort study by Mandeep R. Mehra of the Harvard Medical School and Sapan S. Desai of Surgisphere Corporation, which concluded that the malaria drugs hydroxychloroquine and chloroquine did not improve the condition of COVID-19 patients, and may have harmed some of them.

Where the model says the article warns about this paper:

• In the same section (Surgisphere study on the use of hydroxychloroquine and chloroquine (2020)): The authors of the study then asked The Lancet to retract the article, which was done on 3 June 2020.
• In the same section (Surgisphere study on the use of hydroxychloroquine and chloroquine (2020)): To improve quality control, the editors of The Lancet Group announced changes to the editorial policy in a comment titled "Learning from a retraction" which was published on 22 September 2020.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-23529 / 463899 / #1, read from revision 1372163734 of The Lancet.

The model's reasoning: This directly states the paper's findings about negation and question markers; the Routledge chapter plausibly supports the same claim, and the publisher-error retraction calls for checking the replacement paper.

Other citations on it:

• Hill, Joseph C.; Lillo-Martin, Diane C.; Wood, Sandra K. (2018-12-12). "Homesign systems". Sign Languages. Routledge: 117–133. doi:10.4324/9780429020872-7. ISBN 978-0-429-02087-2. S2CID 239536813.{{cite journal}}: CS1 maint: periodical has ISBN (link)

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-4106 / 2072670 / #1, read from revision 1373558855 of Home sign.

The model's reasoning: This uses the paper's account of headshake negation in homesign, although the causal claim about its replacement is not established by the supplied abstract; check the replacement paper given the publisher-error retraction.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.80. Finding RW-4106 / 2072670 / #2, read from revision 1373558855 of Home sign.

The model's reasoning: Both comparisons repeat the review's conclusions as current evidence; the review was withdrawn as out of date, so an up-to-date source is needed.

The whole list item: Root form implants; the most common type of implant indicated for all uses. Within the root form type of implant, there are roughly 18 variants, all made of titanium but with different shapes and surface textures. There is limited evidence showing that implants with relatively smooth surfaces are less prone to peri-implantitis than implants with rougher surfaces and no evidence showing that any particular type of dental implant has superior long-term success.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-30209 / 1118041 / #1, read from revision 1378363484 of Dental implant.

The model's reasoning: This states the review's conclusion that evidence for dietary efficacy in autism is poor; the other reviews plausibly support it, but the withdrawn review is out of date.

Other citations on it:

• Marí-Bauset S, Zazpe I, Mari-Sanchis A, Llopis-González A, Morales-Suárez-Varela M (December 2014). "Evidence of the gluten-free and casein-free diet in autism spectrum disorders: a systematic review". Journal of Child Neurology. 29 (12): 1718–27. doi:10.1177/0883073814531330. hdl:10171/37087. PMID 24789114.
• Buie T (May 2013). "The relationship of autism and gluten". Clinical Therapeutics (Review). 35 (5): 578–83. doi:10.1016/j.clinthera.2013.04.011. PMID 23688532. At this time, the studies attempting to treat symptoms of autism with diet have not been sufficient to support the general institution of a gluten-free or other diet for all children with autism. ...There may be a subgroup of patients who might benefit from a gluten-free diet, but the symptom or testing profile of these candidates remains unclear.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-30715 / 708662 / #1, read from revision 1363141528 of Gluten-free diet.

The model's reasoning: The quotation presents the withdrawn review's evidence conclusions without a local warning; withdrawal as out of date calls for a current replacement review.

The sentence the footnote ends: However, there is weak evidence to suggest that the use of stabilization splints for the treatment of TMD may be beneficial for reducing pain severity, at rest and on palpation, when compared to no treatment".

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-31677 / 30707 / #7, read from revision 1373270095 of Temporomandibular joint dysfunction.

The model's reasoning: This generalizes the review's finding of insufficient evidence for arthrocentesis and lavage to all TMD treatments; the review does not establish the full claim.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.95. Finding RW-31698 / 30707 / #1, read from revision 1373270095 of Temporomandibular joint dysfunction.

The model's reasoning: The claim appears to draw on the review's adverse-event findings, but broadly worsening symptoms is not established by its reports of mild, transient discomfort or pain.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.77. Finding RW-31698 / 30707 / #3, read from revision 1373270095 of Temporomandibular joint dysfunction.

The model's reasoning: Although the footnote ends at the Ramfjord history sentence, this earlier sentence states the review's conclusion about the absence of evidence linking orthodontic intervention to TMD.

The sentence the footnote ends: A causal relationship between occlusal factors and TMD was championed by Ramfjord in the 1960s.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.88. Finding RW-31674 / 30707 / #3, read from revision 1373270095 of Temporomandibular joint dysfunction.

The model's reasoning: The quotation presents the review's conclusions as evidence; withdrawal as out of date calls for a current replacement source rather than treating the conclusions as disproved.

The sentence the footnote ends: Based on the lack of data, orthodontic treatment cannot be recommended for the treatment or prevention of TMD."

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-31674 / 30707 / #7, read from revision 1373270095 of Temporomandibular joint dysfunction.

The model's reasoning: This repeats the review's pooled finding of CBT benefits in TMD, without its weak-evidence qualification; an editor should check whether this paragraph has independent meta-analysis sources.

The model's confidence: 0.88. Finding RW-31688 / 30707 / #3, read from revision 1373270095 of Temporomandibular joint dysfunction.

The model's reasoning: The breast-cancer miRNA-sponging and downstream protein-regulation claims rest on the retracted LINC00707–miR-876–MTDH study; the other citations plausibly support parts of this combined sentence, including a different breast-cancer pathway.

Other citations on it:

• Jia B, Wang Z, Sun X, Chen J, Zhao J, Qiu X (February 2019). "Long noncoding RNA LINC00707 sponges miR-370-3p to promote osteogenesis of human bone marrow-derived mesenchymal stem cells through upregulating WNT2B". Stem Cell Research & Therapy. 10 (1) 67. doi:10.1186/s13287-019-1161-9. PMC 6387535. PMID 30795799.
• Xie M, Ma T, Xue J, Ma H, Sun M, Zhang Z, et al. (February 2019). "The long intergenic non-protein coding RNA 707 promotes proliferation and metastasis of gastric cancer by interacting with mRNA stabilizing protein HuR". Cancer Letters. 443: 67–79. doi:10.1016/j.canlet.2018.11.032. PMID 30502359. S2CID 54611497.
• Yuan RX, Bao D, Zhang Y (May 2020). "Linc00707 promotes cell proliferation, invasion, and migration via the miR-30c/CTHRC1 regulatory loop in breast cancer". European Review for Medical and Pharmacological Sciences. 24 (9): 4863–4872. doi:10.26355/eurrev_202005_21175. PMID 32432749. S2CID 218759508.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-32587 / 196493 / #1, read from revision 1369045299 of Non-coding RNA.

The model's reasoning: The sentence reports a critical interpretation, but the title alone does not establish whether this was a conclusion of Park's essay; Pekar's own cited criticism plausibly supports it independently.

Other citations on it:

• Pekar 1986, p. 55; Pekar 1990, pp. 32–33. Pekar, Harvey (December 1986). "Maus and Other Topics". The Comics Journal (113). Fantagraphics Books: 54–57. ISSN 0194-7869. Pekar, Harvey (April 1990). "Blood and Thunder". The Comics Journal. 302 (135). Fantagraphics Books: 27–34. Bibcode:1983Natur.302..784D. doi:10.1038/302784a0. ISSN 0194-7869.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.90. Finding RW-34154 / 349114 / #1, read from revision 1378076967 of Maus.

The model's reasoning: The recommendation of flossing relies on the review's finding that adding flossing to brushing reduces gingivitis; the other oral-health source plausibly supports this advice, but the review was withdrawn as out of date.

Other citations on it:

• "How to Get Rid of Puffy Gums with Braces". Orthodontics Australia. 2021-05-02. Retrieved 2021-07-21.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-30653 / 19637986 / #1, read from revision 1378039819 of Gingivitis.

The model's reasoning: Including flossing as a treatment repeats the review's finding of reduced gingivitis and explicitly draws on the prevention advice that cites it; an editor should check whether this paragraph has an independent source.

Where the model found it: The focus of treatment is to remove plaque. Therapy is aimed at the reduction of oral bacteria and may take the form of regular periodic visits to a dental professional, together with adequate oral hygiene home care. Thus, several of the methods used in the prevention of gingivitis can also be used for the treatment of manifest gingivitis, such as scaling, root planing, curettage, mouth washes containing chlorhexidine or hydrogen peroxide, and flossing. Interdental brushes also help remove any causative agents.

The model's confidence: 0.83. Finding RW-30653 / 19637986 / #2, read from revision 1378039819 of Gingivitis.

The model's reasoning: The sentence attributes a fatigue benefit to the cited review, but its title alone does not establish this amantadine finding, and the stated year differs from the supplied paper's year.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.90. Finding RW-30944 / 1128227 / #1, read from revision 1369113052 of Amantadine.

The model's reasoning: The sentence presents an effectiveness finding within the review's scope as established; the review was withdrawn as out of date and replaced.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.96. Finding RW-30651 / 147728 / #1, read from revision 1377234121 of Toothbrush.

The model's reasoning: This clause reports the cited review's conclusion about comparison with floss; the other source plausibly supports greater plaque removal, but not this insufficient-evidence conclusion.

The sentence the footnote ends: Although there is some evidence that after tooth brushing with a conventional tooth brush, interdental brushes remove more plaque than dental floss, a systematic review reported insufficient evidence to determine such an association.

Other citations on it:

• Gluch, Joan I. (2012). "As an Adjunct to Tooth Brushing, Interdental Brushes (IDBs) are More Effective in Removing Plaque as Compared with Brushing Alone or the Combination Use of Tooth Brushing and Dental Floss". Journal of Evidence Based Dental Practice. 12 (2): 81–83. doi:10.1016/j.jebdp.2012.03.016. ISSN 1532-3382. PMID 22726785.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.90. Finding RW-30651 / 147728 / #2, read from revision 1377234121 of Toothbrush.

The model's reasoning: The claim about touch therapy uses the review's finding of no robust evidence for acute-wound healing, but generalizes it to all health benefits; other cited sources plausibly support the broader assessment.

The sentence the footnote ends: Null has been also a supporter of touch therapy and magnet therapy, both of which have been long determined to not provide any tangible health benefits.

Other citations on it:

• Williams, Robyn (2000). Scary Monsters and Bright Ideas. UNSW Press. p. 38. ISBN 978-0-86840-707-4.
• Rosa, Linda; Rosa, E; Sarner, L; Barrett, S (April 1, 1998). "A Close Look at Therapeutic Touch". JAMA. 279 (13): 1005–1010. doi:10.1001/jama.279.13.1005. PMID 9533499.
• Ades TB, ed. (2009). "Therapeutic Touch". American Cancer Society Complete Guide to Complementary and Alternative Cancer Therapies (2nd ed.). American Cancer Society. pp. 248–251. ISBN 978-0-944235-71-3.
• Trick or Treatment. Corgi. 2008. pp. 267–268.
• Russell, Jill; Rovere, Amy, eds. (2009). "Magnetic therapy". American Cancer Society Complete Guide to Complementary & Alternative Cancer Therapies (2nd ed.). Atlanta, Ga.: American Cancer Society. ISBN 978-0-944235-71-3. See archived online version "Magnetic Therapy". November 1, 2008. Archived from the original on November 12, 2012. Electromagnetic therapy is a related field. See chapter in ACS book just referenced, and archived ACS webpage on that.
• Park, Robert L. (2000). Voodoo Science: The Road from Foolishness to Fraud. New York, New York: Oxford University Press. pp. 58–63. ISBN 0-19-513515-6. Not only are magnetic fields of no value in healing, you might characterize these as "homeopathic" magnetic fields.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-31159 / 3346025 / #1, read from revision 1340347277 of Gary Null.

The model's reasoning: This characterization draws on the review's tolerability findings, although it is broader than its cautious conclusion about fewer adverse-event dropouts; the outdated review needs replacement.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.92. Finding RW-31029 / 2038711 / #1, read from revision 1370866766 of Trimetazidine.

The model's reasoning: Both sentences reproduce the review's evidence-gap assessment and call for long-term comparative trials; withdrawal as out of date calls for a current source.

The sentence the footnote ends: Long-term randomized, controlled trials comparing trimetazidine against standard antianginal agents using clinically important outcomes would be justifiable.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-31029 / 2038711 / #2, read from revision 1370866766 of Trimetazidine.

The model's reasoning: The exercise-ischemia, angina-frequency and nitrate-use benefits repeat the review's results, but the paragraph includes additional claims and may have independent sources whose footnotes an editor must check.

The model's confidence: 0.85. Finding RW-31029 / 2038711 / #3, read from revision 1370866766 of Trimetazidine.

The model's reasoning: This repeats the withdrawn review's conclusion that poor-quality evidence prevents confident assessments, although its broader scope suggests other reviews may support it; an editor should check the paragraph's own sources.

Where the model found it: Transcutaneous electrical nerve stimulation is a commonly used treatment approach to alleviate acute and chronic pain by reducing the sensitization of dorsal horn neurons, elevating levels of gamma-aminobutyric acid and glycine, and inhibiting glial activation. Many systematic reviews and meta-analyses assessing clinical trials looking at the efficacy of TENS for different sources of pain, however, have been inconclusive due to a lack of high-quality and unbiased evidence. Potential benefits of TENS treatment include its safety profile, relative affordability, ease of self-administration, and availability over-the-counter without a prescription. In principle, an adequate intensity of stimulation is necessary to achieve pain relief with TENS. An analysis of treatment fidelity—meaning that the delivery of TENS in a trial was in accordance with current clinical advice, such as using "a strong but comfortable sensation" and suitable, frequent treatment durations—showed that higher-fidelity trials tended to have a positive outcome. Repeated use of TENS at the same intensity and frequency settings may lead to analgesic tolerance, this effect may be reduced by varying stimulation frequency or increasing intensity. Recent clinical trials indicate that TENS may only be effective for a subset of patients, with approximately 44% of individuals experience a meaningful reduction in pain. Compared to pharmacological drug treatments, TENS can have a similar effectiveness in some patients while presenting a lower risk of adverse effects.

The model's confidence: 0.70. Finding RW-32379 / 683583 / #2, read from revision 1372039004 of Transcutaneous electrical nerve stimulation.

The model's reasoning: This reports the withdrawn review's conclusions about gingivitis and plaque without noting its withdrawal; it was withdrawn as out of date and replaced, rather than for invalid findings.

The reference isn't tagged {{Retracted}}.

The model's confidence: 1.00. Finding RW-30653 / 487691 / #1, read from revision 1376165893 of Dental floss.

The model's reasoning: This states the paper's central Ash1 recruitment and Ubx activation finding as fact, with no other citation; the fabrication findings undermine that evidence.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-500 / 16770101 / #2, read from revision 1378516518 of Long non-coding RNA.

The model's reasoning: The review concluded that talc was the most effective sclerosant; the accompanying 2018 treatment review plausibly supports this comparison, and the retracted review was replaced.

The sentence the footnote ends: However, for scarring down the pleura, talc appears to be the better option although indwelling pleural catheters are at least as effective in reducing the symptoms of an effusion(such as dyspnea).

Other citations on it:

• Thomas R, Murray K, Lee YC (April 2018). "Treatment Approaches for Malignant Pleural Effusion". JAMA. 319 (14): 1507–1508. doi:10.1001/jama.2018.1323. PMID 29634827.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-32770 / 1010309 / #1, read from revision 1376179398 of Bleomycin.

The model's reasoning: The lead repeats the talc-superiority conclusion cited to the retracted review under Other uses; an editor should check whether the lead has independent supporting sources.

Where the model found it: Bleomycin is a medication primarily used to treat cancer. This includes Hodgkin's lymphoma, non-Hodgkin's lymphoma, testicular cancer, ovarian cancer, and cervical cancer among others. Typically used with other cancer medications, it can be given intravenously, by injection into a muscle or under the skin. It may also be administered inside the chest to help prevent the recurrence of a pleural effusion due to cancer; however, talc is better for this. It may sometimes be used to treat other difficult-to-treat skin lesions such as plantar warts in immunocompromised patients.

The model's confidence: 0.97. Finding RW-32770 / 1010309 / #2, read from revision 1376179398 of Bleomycin.

The model's reasoning: This reports the paper's claimed detection of XMRV without a retraction warning in this paragraph; the study is notable in the controversy, but its results were unreliable.

Where the model says the article warns about this paper:

• In the same section (XMRV retrovirus): XMRV, a virus claimed to be associated with CFS in a paper that was later retracted
• In the same section (XMRV retrovirus): In 2011, the editor of Science formally retracted its XMRV paper while the Proceedings of the National Academy of Sciences similarly retracted a 2010 paper which had appeared to support the finding of a connection between XMRV and CFS.
• In the same section (XMRV retrovirus): Studies eventually concluded that neither people nor the blood supply had been infected with the XMRV virus, and the origin of the virus was likely a lab contaminant in the supplies used by the polymerase chain reaction (PCR) process of the studies that found virus in blood.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-492 / 16824033 / #1, read from revision 1376392525 of Controversies related to ME/CFS.

The model's reasoning: The fitness and body-composition claims overlap the paper's main findings, although the mental and social claims do not; an editor must check whether this lead summary has independent sources.

Where the model found it: Studies indicate that CrossFit can have positive effects on a number of physical fitness parameters and body composition, as well as on the mental state and social life of its participants. CrossFit has been criticized for causing more injuries than other sporting activities such as weightlifting, although a review article in the Journal of Sports Rehabilitation found that "the risk of injury from participation in CrossFit is comparable to or lower than some common forms of exercise or strength training". Its health benefits and injury rates are determined to be similar to other exercise programs. There are also concerns that its methodology may cause exertional rhabdomyolysis, a possible life-threatening condition also found in other sports, resulting from a breakdown of muscle from extreme exertion.

Where the model says the article warns about this paper:

• Later (Lawsuit by CrossFit, Inc. against the NSCA): In an erratum later published by the journal, it is stated that "after the article was published, 10 of the 11 participants who did not complete the study have provided their reasons for not finishing, with only 2 mentioning injury or health conditions that prevented them from completing follow-up testing."
• Later (Lawsuit by CrossFit, Inc. against the NSCA): The NSCA denies CrossFit, Inc.'s allegations but issued an erratum acknowledging that the injury data were incorrect.
• Later (Lawsuit by CrossFit, Inc. against the NSCA): In September 2016, the District Court ruled in favor of CrossFit because the injury data were found to be false, but not that the NSCA was commercially motivated or that publishing the study was defamatory, as the NSCA no longer stood behind the study. The Journal of Strength and Conditioning Research eventually retracted the paper in its July 2017 issue.
• Later (Lawsuit by CrossFit, Inc. against the NSCA): In May 2017, the Court issued 17 issues sanctions against the NSCA, writing that the organization did have a commercial motive to falsify the data, had published the false data knowingly to disparage CrossFit, and had misled the public with their erratum.

The model's confidence: 0.90. Finding RW-8161 / 11568969 / #2, read from revision 1375365469 of CrossFit.

The model's reasoning: The text reports the opioid mechanism discussed in the review without noting its withdrawal; the other review plausibly supports this account of the hypothesis.

Other citations on it:

• Shattock P, Whiteley P (April 2002). "Biochemical aspects in autism spectrum disorders: updating the opioid-excess theory and presenting new opportunities for biomedical intervention". Expert Opinion on Therapeutic Targets (Review). 6 (2): 175–183. doi:10.1517/14728222.6.2.175. PMID 12223079. S2CID 40904799.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.87. Finding RW-30715 / 4071231 / #2, read from revision 1376296858 of Causes of autism.

The model's reasoning: This reports abnormal urinary peptide measurements summarized in the withdrawn review, with no other citation on the sentence.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-30715 / 4071231 / #3, read from revision 1376296858 of Causes of autism.

The model's reasoning: This reports cerebrospinal-fluid measurements summarized in the withdrawn review; withdrawal as out of date calls for an updated source rather than automatic removal.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-30715 / 4071231 / #4, read from revision 1376296858 of Causes of autism.

The model's reasoning: This repeats the food-derived opioid mechanism supported by the withdrawn review further down; an editor must check whether this paragraph has independent support.

Where the model found it: The "leaky gut syndrome" hypothesis developed by Andrew Wakefield, known for his fraudulent study on another cause of autism, is popular among parents of children with autism. It is based on the idea that defects in the intestinal barrier produce an excessive increase in intestinal permeability, allowing substances present in the intestine (including bacteria, environmental toxins, and food antigens) to pass into the blood. The data supporting this theory are limited and contradictory, since both increased intestinal permeability and normal permeability have been documented in people with autism. Studies with mice provide some support to this theory and suggest the importance of intestinal flora, demonstrating that the normalization of the intestinal barrier was associated with an improvement in some of the autism-like behaviors. Studies on subgroups of people with autism showed the presence of high plasma levels of zonulin, a protein that regulates permeability opening the "pores" of the intestinal wall, as well as intestinal dysbiosis (reduced levels of Bifidobacteria and increased abundance of Akkermansia muciniphila, Escherichia coli, Clostridia and Candida fungi that promote the production of proinflammatory cytokines, all of which produces excessive intestinal permeability. This allows passage of bacterial endotoxins from the gut into the bloodstream, stimulating liver cells to secrete tumor necrosis factor alpha (TNFα), which modulates blood–brain barrier permeability. Studies on ASD people showed that TNFα cascades produce proinflammatory cytokines, leading to peripheral inflammation and activation of microglia in the brain, which indicates neuroinflammation. In addition, neuroactive opioid peptides from digested foods have been shown to leak into the bloodstream and permeate the blood–brain barrier, influencing neural cells and causing autistic symptoms. (See Endogenous opiate precursor theory)

The model's confidence: 0.80. Finding RW-30715 / 4071231 / #5, read from revision 1376296858 of Causes of autism.

The model's reasoning: This presents the paper's redox–gut anaerobe association as established; the separate retraction note does not qualify the claim within this paragraph.

Where the model says the article warns about this paper:

• In a note (Notes): This study was later retracted over ethical oversight concerns.

Other citations on it:

• Million, Matthieu; Raoult, Didier (December 2018). "Linking gut redox to human microbiome". Human Microbiome Journal. 10: 27–32. doi:10.1016/j.humic.2018.07.002.
• This study was later retracted over ethical oversight concerns.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.98. Finding RW-48648 / 92640 / #1, read from revision 1378083612 of Anaerobic organism.

The model's reasoning: The paragraph retells the paper's findings without locally identifying them as retracted or falsified; saying the association was not proved is not a retraction warning.

The sentence the footnote ends: In February 1998, a group led by Andrew Wakefield published a paper in the British medical journal The Lancet, supported by a press conference at the Royal Free Hospital in London, where the research was carried out.

Where the model says the article warns about this paper:

• In the opening paragraphs: Fraudulent Lancet MMR vaccine-autism study
• In the opening paragraphs: The paper falsely claimed causative links between the measles, mumps, and rubella (MMR) vaccine and colitis, and between colitis and autism. The fraud involved data bias and manipulation, and two undisclosed conflicts of interest. It was exposed in a lengthy Sunday Times investigation by reporter Brian Deer, resulting in the paper's retraction in February 2010 and Wakefield's being discredited and struck off the UK medical register three months later. In the paper, Wakefield fabricated evidence to suggest a new "syndrome" existed, which he called "autistic enterocolitis".
• In the opening paragraphs: a fraudulent research paper
• Later (Conflict of interest): Based on Deer's evidence, the Lancet's editor-in-chief Richard Horton said Wakefield's paper should have never been published because its findings were "entirely flawed".
• Later (No ethical approval): The dispute over this would remain unresolved, however, until settled in the English High Court in March 2012, where a senior judge vindicated Deer. Quoting the text, Justice Mitting ruled, "This statement was untrue and should not have been included in the paper."
• Later (Retraction of an interpretation): Retraction of an interpretation
• Later (Retraction of an interpretation): In March 2004, immediately following the news of the conflict of interest allegations, ten of Wakefield's 12 coauthors retracted this interpretation, while insisting that the possibility of a distinctive gastrointestinal condition in children with autism merited further investigation.
• Later (Retraction of an interpretation): The Lancet paper was partially retracted in 2004 and fully retracted in 2010, when Lancet's editor-in-chief Richard Horton described it as "utterly false" and said that the journal had been deceived.
• Later (Retraction of an interpretation): In 2011, Deer provided further information on Wakefield's improper research practices to the British Medical Journal, which in a signed editorial described the original paper as fraudulent.
• Later (Manipulation of data): The Sunday Times continued the investigation, and on 8 February 2009, Brian Deer reported that Wakefield had "fixed" results and "manipulated" patient data in the Lancet, creating the appearance of a link with autism. Wakefield falsely denied these allegations, and even filed a complaint with the Press Complaints Commission (PCC) over this article on 13 March 2009.
• Later (UK General Medical Council inquiry): In response to the GMC investigation and findings, the editors of the Lancet announced on 2 February 2010 that they "fully retract this paper from the published record". The Lancet's editor-in-chief Richard Horton described it as "utterly false" and said that the journal had been deceived.
• Later (Full retraction and fraud revelations): Full retraction and fraud revelations
• Later (Full retraction and fraud revelations): In an April 2010 report in The BMJ, Deer expanded on the laboratory aspects of his findings recounting how normal clinical histopathology results generated by the Royal Free Hospital were later changed in the medical school to abnormal results, published in the Lancet.
• Later (Full retraction and fraud revelations): On 5 January 2011, The BMJ published the first of a series of articles by Brian Deer, detailing how Wakefield and his colleagues had faked some of the data behind the 1998 Lancet article. By looking at the records and interviewing the parents, Deer found that for all 12 children in the Wakefield study, diagnoses had been tweaked or dates changed to fit the article's conclusion.
• Later (Full retraction and fraud revelations): In an editorial accompanying Deer's 2011 series, The BMJ said, "it has taken the diligent scepticism of one man, standing outside medicine and science, to show that the paper was in fact an elaborate fraud", and asked:
• Later (Full retraction and fraud revelations): A great deal of thought and effort must have gone into drafting the paper to achieve the results he wanted: the discrepancies all led in one direction; misreporting was gross.
• Later (Full retraction and fraud revelations): The Lancet paper was a case series of 12 child patients; it reported a proposed "new syndrome" of enterocolitis and regressive autism and associated this with MMR as an "apparent precipitating event". But in fact:
• Later (Full retraction and fraud revelations): Three of nine children reported with regressive autism did not have autism diagnosed at all. Only one child clearly had regressive autism.
• Later (Full retraction and fraud revelations): Despite the paper claiming that all 12 children were "previously normal", five had documented pre-existing developmental concerns.
• Later (Full retraction and fraud revelations): Some children were reported to have experienced first behavioural symptoms within days of MMR, but the records documented these as starting some months after vaccination.
• Later (Full retraction and fraud revelations): In nine cases, unremarkable colonic histopathology results—noting no or minimal fluctuations in inflammatory cell populations—were changed after a medical school "research review" to "non-specific colitis".
• Later (Full retraction and fraud revelations): The parents of eight children were reported as blaming MMR, but 11 families made this allegation at the hospital. The exclusion of three allegations—all giving times to onset of problems in months—helped to create the appearance of a 14 day temporal link.
• Later (Full retraction and fraud revelations): In subsequent disclosures from the investigation, Deer obtained copies of unpublished gastrointestinal pathology reports on the children in the Lancet study that Wakefield had claimed showed "non-specific colitis" and "autistic enterocolitis". But expert analyses of these reports found bowel biopsies from the children to be overwhelmingly normal and with no evidence of any enterocolitis at all.
• Later (Full retraction and fraud revelations): One described the paper as "fraudulent" while another complained of "outright fabrication".
• Later (Aftermath): Promotion of the claimed link, which continues in anti-vaccination propaganda despite being refuted, led to an increase in the incidence of measles and mumps, resulting in deaths and serious permanent injuries.
• Later (Aftermath): Among commentators drawing on Deer's investigation, academic Peter N. Steinmetz summarizes six fabrications and falsifications in the paper itself and in Wakefield's response: findings of non-specific colitis; timing of MMR vaccine administration and first behavioral symptoms; findings of regressive autism; ethics consent statement; conflict of interest statement; and methods of patient referral.

Other citations on it:

• Moore Andrew (2006). "Bad science in the headlines: Who takes responsibility when science is distorted in the mass media?". EMBO Reports. 7 (12): 1193–1196. doi:10.1038/sj.embor.7400862. PMC 1794697. PMID 17139292.
• "New Research Links Autism and Bowel Disease" (PDF). Royal Free Hospital School of Medicine (Press release). 26 February 1998. Archived (PDF) from the original on 23 March 2011. Retrieved 16 November 2022 – via briandeer.com.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-4036 / 59849415 / #2, read from revision 1377289380 of Wakefield MMR-autism fraud.

The model's reasoning: The citation can cover the preceding sentences; this opening claim states the epithelial–macrophage interaction underlying the paper's homeostatic circuit, whereas the footnote-ending sentence is general cytokine background.

The sentence the footnote ends: TGF-β is a multifunctional cytokine that modulates a variety of biological processes such as cell growth, apoptosis, extracellular matrix synthesis, inflammation, and immune responses.

The reference is tagged {{Retracted}}.

The model's confidence: 0.91. Finding RW-2268 / 8129870 / #1, read from revision 1329563606 of Alveolar macrophage.

The model's reasoning: In this paragraph, macrophage suppression is presented as the inhibitory component of the lung circuit described in the abstract.

The reference is tagged {{Retracted}}.

The model's confidence: 0.90. Finding RW-2268 / 8129870 / #4, read from revision 1329563606 of Alveolar macrophage.

The model's reasoning: The alveolar actin response introduces the microbial-triggered mechanism that releases macrophages from inhibition in the paper's circuit.

The reference is tagged {{Retracted}}.

The model's confidence: 0.88. Finding RW-2268 / 8129870 / #6, read from revision 1329563606 of Alveolar macrophage.

The model's reasoning: This directly states the paper's proposed mechanism for transiently circumventing epithelial TGFβ-mediated inhibition.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-2268 / 8129870 / #7, read from revision 1329563606 of Alveolar macrophage.

The model's reasoning: In this paragraph, activation refers to the preceding paper-derived release from epithelial inhibition, making this the functional outcome of that mechanism.

The reference is tagged {{Retracted}}.

The model's confidence: 0.84. Finding RW-2268 / 8129870 / #8, read from revision 1329563606 of Alveolar macrophage.

The model's reasoning: The sentence presents macrophage metalloproteinase production as the initiating step of the paper's proposed restoration of tonic inhibition.

The reference is tagged {{Retracted}}.

The model's confidence: 0.92. Finding RW-2268 / 8129870 / #9, read from revision 1329563606 of Alveolar macrophage.

The model's reasoning: This directly repeats the circuit's restoration mechanism; the contemporaneous commentary may depend on the retracted study, and the asthma review's support for the complete circuit is uncertain.

Other citations on it:

• Lambrecht, BN (April 2006). ""Alveolar macrophage in the driver's seat"". Immunity. 24 (4): 366–8. doi:10.1016/j.immuni.2006.03.008. PMID 16618595.
• Ohbayashi H, Shimokata K (April 2005). "Matrix metalloproteinase-9 and airway remodeling in asthma". Current Drug Targets. Inflammation and Allergy. 4 (2): 177–81. doi:10.2174/1568010053586246. PMID 15853739.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-2268 / 8129870 / #10, read from revision 1329563606 of Alveolar macrophage.

The model's reasoning: Both the efficacy claim and uncertainty about comparative efficacy reflect the retracted review's findings; no other source is cited.

The sentence the footnote ends: It is unclear if it is more or less effective than other agents such as droperidol, metoclopramide, or ondansetron.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-30918 / 429616 / #1, read from revision 1335737669 of Granisetron.

The model's reasoning: The discovery and roughly three-second duration are results reported in the retracted paper; the CNN report plausibly repeats them but may rely on the same study.

Other citations on it:

• Strickland, Ashley (13 July 2022). "Mysterious fast radio burst in space has a 'heartbeat' pattern". CNN. Retrieved 22 August 2022.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-72832 / 62720662 / #1, read from revision 1378291880 of 2022 in science.

The model's reasoning: This echoes the review's finding of possible benefit from cognitive behavioural therapy, although its pooled trials concerned temporomandibular disorder; an editor should check whether this paragraph has independent, BMS-specific support.

Where the model found it: Other treatments which have been used include atypical antipsychotics, histamine receptor antagonists, and dopamine agonists. Supplementation with vitamin complexes and cognitive behavioral therapy may be helpful in the management of burning mouth syndrome.

The model's confidence: 0.72. Finding RW-31688 / 5186108 / #2, read from revision 1375134929 of Burning mouth syndrome.

The model's reasoning: This reports the paper's claimed derivation of cloned human stem cells without telling readers in this paragraph that the result was fabricated or retracted.

Where the model says the article warns about this paper:

• In the opening paragraphs: However, he became infamous around November 2005 for fabricating a series of stem cell experiments that were published in high-profile journals in a case known as the Hwang affair.
• In the opening paragraphs: Shortly after this, data from his human cloning experiments was revealed to have been falsified.
• In the opening paragraphs: On May 12, 2006, Hwang was charged with embezzlement and bioethics law violations after it emerged much of his stem cell research had been faked.
• In the opening paragraphs: On the same day, CNN reported that the scientist in 2006 had admitted faking his findings after questions of impropriety had emerged.
• In the same section (Timeline): The university announced on January 10, 2006, that Hwang's 2004 and 2005 papers in Science were both fabricated. Following on the confirmation of scientific misconduct, on January 11, Science unconditionally retracted both of Hwang's papers.
• Later (Official probe and confirmation of fraud): Official probe and confirmation of fraud
• Later (Official probe and confirmation of fraud): The panel stated that Hwang's 2004 Science paper was also fabricated
• Later (Official probe and confirmation of fraud): The panel, in conclusion, stated that Hwang's team intentionally fabricated the data in both the 2004 and the 2005 papers, as described by Myung Hee Chung (Head of Seoul National University's investigation) and that it was an act of "deception of the scientific community and the public at large".
• Later (Parthenogenesis): However, further examination of the cells' chromosomes shows the same indicators of parthenogenesis in those extracted stem cells as are found in the mice created by Tokyo scientists in 2004. Although Hwang deceived the world about being the first to create artificially cloned human embryos, he did contribute a major breakthrough to the field of stem cell research.
• Later (Parthenogenesis): when the fabricated cloning results were announced
• Later (Supporting lawmakers group): After most of Hwang's claims were confirmed to be false on January 10, 2006, some lawmakers revealed that Hwang had made several campaign donations to them and other lawmakers.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-1004 / 1962277 / #1, read from revision 1376690909 of Hwang Woo-suk.

The model's reasoning: This comparison repeats the 2004 paper's claimed stem-cell derivation as fact; any independent support in this paragraph needs checking.

Where the model found it: Hwang's team announced an even greater achievement a year later in May 2005, and claimed they had created 11 human embryonic stem cells using 185 eggs. His work, published in the June 17 issue of Science, was instantly hailed as a breakthrough in biotechnology because the cells were allegedly created with somatic cells from patients of different age and gender, while the stem cell of 2004 was created with eggs and somatic cells from a single female donor. This meant every patient could receive custom-made treatment with no immune reactions. In addition, Hwang's claim meant that his team had boosted their success rate by 14 times and that this technology could be medically viable.

Where the model says the article warns about this paper:

• In the opening paragraphs: However, he became infamous around November 2005 for fabricating a series of stem cell experiments that were published in high-profile journals in a case known as the Hwang affair.
• In the opening paragraphs: Shortly after this, data from his human cloning experiments was revealed to have been falsified.
• In the opening paragraphs: On May 12, 2006, Hwang was charged with embezzlement and bioethics law violations after it emerged much of his stem cell research had been faked.
• In the opening paragraphs: On the same day, CNN reported that the scientist in 2006 had admitted faking his findings after questions of impropriety had emerged.
• In the same section (Timeline): The university announced on January 10, 2006, that Hwang's 2004 and 2005 papers in Science were both fabricated. Following on the confirmation of scientific misconduct, on January 11, Science unconditionally retracted both of Hwang's papers.
• Later (Official probe and confirmation of fraud): Official probe and confirmation of fraud
• Later (Official probe and confirmation of fraud): The panel stated that Hwang's 2004 Science paper was also fabricated
• Later (Official probe and confirmation of fraud): The panel, in conclusion, stated that Hwang's team intentionally fabricated the data in both the 2004 and the 2005 papers, as described by Myung Hee Chung (Head of Seoul National University's investigation) and that it was an act of "deception of the scientific community and the public at large".
• Later (Parthenogenesis): However, further examination of the cells' chromosomes shows the same indicators of parthenogenesis in those extracted stem cells as are found in the mice created by Tokyo scientists in 2004. Although Hwang deceived the world about being the first to create artificially cloned human embryos, he did contribute a major breakthrough to the field of stem cell research.
• Later (Parthenogenesis): when the fabricated cloning results were announced
• Later (Supporting lawmakers group): After most of Hwang's claims were confirmed to be false on January 10, 2006, some lawmakers revealed that Hwang had made several campaign donations to them and other lawmakers.

The model's confidence: 0.97. Finding RW-1004 / 1962277 / #3, read from revision 1376690909 of Hwang Woo-suk.

The model's reasoning: The quoted endorsement treats the 2004 cloning claim as established; Time supplies the historical quotation, not independent validation of the result, and local sources need checking.

Where the model found it: Hwang made further headlines in May 2005 when he criticized U.S. President George W. Bush's policy on embryonic stem cell research. Also, Time magazine named Hwang one of its "People Who Mattered 2004", stating that Hwang "has already proved that human cloning is no longer science fiction, but a fact of life."

Where the model says the article warns about this paper:

• In the opening paragraphs: However, he became infamous around November 2005 for fabricating a series of stem cell experiments that were published in high-profile journals in a case known as the Hwang affair.
• In the opening paragraphs: Shortly after this, data from his human cloning experiments was revealed to have been falsified.
• In the opening paragraphs: On May 12, 2006, Hwang was charged with embezzlement and bioethics law violations after it emerged much of his stem cell research had been faked.
• In the opening paragraphs: On the same day, CNN reported that the scientist in 2006 had admitted faking his findings after questions of impropriety had emerged.
• In the same section (Timeline): The university announced on January 10, 2006, that Hwang's 2004 and 2005 papers in Science were both fabricated. Following on the confirmation of scientific misconduct, on January 11, Science unconditionally retracted both of Hwang's papers.
• Later (Official probe and confirmation of fraud): Official probe and confirmation of fraud
• Later (Official probe and confirmation of fraud): The panel stated that Hwang's 2004 Science paper was also fabricated
• Later (Official probe and confirmation of fraud): The panel, in conclusion, stated that Hwang's team intentionally fabricated the data in both the 2004 and the 2005 papers, as described by Myung Hee Chung (Head of Seoul National University's investigation) and that it was an act of "deception of the scientific community and the public at large".
• Later (Parthenogenesis): However, further examination of the cells' chromosomes shows the same indicators of parthenogenesis in those extracted stem cells as are found in the mice created by Tokyo scientists in 2004. Although Hwang deceived the world about being the first to create artificially cloned human embryos, he did contribute a major breakthrough to the field of stem cell research.
• Later (Parthenogenesis): when the fabricated cloning results were announced
• Later (Supporting lawmakers group): After most of Hwang's claims were confirmed to be false on January 10, 2006, some lawmakers revealed that Hwang had made several campaign donations to them and other lawmakers.

The model's confidence: 0.93. Finding RW-1004 / 1962277 / #4, read from revision 1376690909 of Hwang Woo-suk.

The model's reasoning: The introductory clause treats the earlier human-cloning achievement underlying the 2004 paper as fact; the paragraph questions later work but does not refute this clause, so local sources need checking.

Where the model found it: His team, which cloned the first human embryo to use for research, said they had used the same technology to create batches of embryonic stem cells from nine patients. According to Hwang, the result was much more efficient than they had hoped. Hwang's integrity as a researcher was again put in doubt when it was revealed that PD Su-cheop had scheduled a follow-up report questioning his achievement published in Science in June 2005, which stated he had cloned 11 lines of embryonic stem cells. This caused furious backlash among many South Koreans, and the reaction only intensified when it was discovered that Kim Sun-Jong, one of Hwang's researchers from MizMedi, had been coerced by illegal means to testify against Hwang. As a result, the scheduled broadcast was canceled and the network made a public apology to the nation, everyone more or less operating under the assumption that the show was at fault and not Hwang. Yet, other news outlets began to question Hwang's claims.

Where the model says the article warns about this paper:

• In the opening paragraphs: However, he became infamous around November 2005 for fabricating a series of stem cell experiments that were published in high-profile journals in a case known as the Hwang affair.
• In the opening paragraphs: Shortly after this, data from his human cloning experiments was revealed to have been falsified.
• In the opening paragraphs: On May 12, 2006, Hwang was charged with embezzlement and bioethics law violations after it emerged much of his stem cell research had been faked.
• In the opening paragraphs: On the same day, CNN reported that the scientist in 2006 had admitted faking his findings after questions of impropriety had emerged.
• Earlier (Timeline): The university announced on January 10, 2006, that Hwang's 2004 and 2005 papers in Science were both fabricated. Following on the confirmation of scientific misconduct, on January 11, Science unconditionally retracted both of Hwang's papers.
• In the same section (Official probe and confirmation of fraud): Official probe and confirmation of fraud
• In the same section (Official probe and confirmation of fraud): The panel stated that Hwang's 2004 Science paper was also fabricated
• In the same section (Official probe and confirmation of fraud): The panel, in conclusion, stated that Hwang's team intentionally fabricated the data in both the 2004 and the 2005 papers, as described by Myung Hee Chung (Head of Seoul National University's investigation) and that it was an act of "deception of the scientific community and the public at large".
• Later (Parthenogenesis): However, further examination of the cells' chromosomes shows the same indicators of parthenogenesis in those extracted stem cells as are found in the mice created by Tokyo scientists in 2004. Although Hwang deceived the world about being the first to create artificially cloned human embryos, he did contribute a major breakthrough to the field of stem cell research.
• Later (Parthenogenesis): when the fabricated cloning results were announced
• Later (Supporting lawmakers group): After most of Hwang's claims were confirmed to be false on January 10, 2006, some lawmakers revealed that Hwang had made several campaign donations to them and other lawmakers.

The model's confidence: 0.94. Finding RW-1004 / 1962277 / #5, read from revision 1376690909 of Hwang Woo-suk.

The model's reasoning: This retells the 2004 finding with expressions of doubt, but does not state that it was refuted or retracted; sources for the historical account need checking.

Where the model found it: Some scientists have started questioning Hwang's earlier work published in Science in February 2004, in which he claimed to have cloned embryonic stem cells. Maria Biotech head Park Se-pill said, "Up until now, I have believed Hwang did derive cloned embryonic stem cells although he admitted to misconduct in his follow-up paper on patient-specific stem cells...Now, I am not sure whether the cloned stem cell really existed."

Where the model says the article warns about this paper:

• In the opening paragraphs: However, he became infamous around November 2005 for fabricating a series of stem cell experiments that were published in high-profile journals in a case known as the Hwang affair.
• In the opening paragraphs: Shortly after this, data from his human cloning experiments was revealed to have been falsified.
• In the opening paragraphs: On May 12, 2006, Hwang was charged with embezzlement and bioethics law violations after it emerged much of his stem cell research had been faked.
• In the opening paragraphs: On the same day, CNN reported that the scientist in 2006 had admitted faking his findings after questions of impropriety had emerged.
• Earlier (Timeline): The university announced on January 10, 2006, that Hwang's 2004 and 2005 papers in Science were both fabricated. Following on the confirmation of scientific misconduct, on January 11, Science unconditionally retracted both of Hwang's papers.
• In the same section (Official probe and confirmation of fraud): Official probe and confirmation of fraud
• In the same section (Official probe and confirmation of fraud): The panel stated that Hwang's 2004 Science paper was also fabricated
• In the same section (Official probe and confirmation of fraud): The panel, in conclusion, stated that Hwang's team intentionally fabricated the data in both the 2004 and the 2005 papers, as described by Myung Hee Chung (Head of Seoul National University's investigation) and that it was an act of "deception of the scientific community and the public at large".
• Later (Parthenogenesis): However, further examination of the cells' chromosomes shows the same indicators of parthenogenesis in those extracted stem cells as are found in the mice created by Tokyo scientists in 2004. Although Hwang deceived the world about being the first to create artificially cloned human embryos, he did contribute a major breakthrough to the field of stem cell research.
• Later (Parthenogenesis): when the fabricated cloning results were announced
• Later (Supporting lawmakers group): After most of Hwang's claims were confirmed to be false on January 10, 2006, some lawmakers revealed that Hwang had made several campaign donations to them and other lawmakers.

The model's confidence: 0.86. Finding RW-1004 / 1962277 / #6, read from revision 1376690909 of Hwang Woo-suk.

The model's reasoning: The paragraph uses the fabricated patient-specific cell lines as evidence for immune compatibility, improved efficiency and medical viability, despite describing the initial results as alleged.

The sentence the footnote ends: His work, published in the June 17 issue of Science, was instantly hailed as a breakthrough in biotechnology because the cells were allegedly created with somatic cells from patients of different age and gender, while the stem cell of 2004 was created with eggs and somatic cells from a single female donor.

Where the model says the article warns about this paper:

• In the opening paragraphs: However, he became infamous around November 2005 for fabricating a series of stem cell experiments that were published in high-profile journals in a case known as the Hwang affair.
• In the opening paragraphs: Shortly after this, data from his human cloning experiments was revealed to have been falsified.
• In the opening paragraphs: On May 12, 2006, Hwang was charged with embezzlement and bioethics law violations after it emerged much of his stem cell research had been faked.
• In the opening paragraphs: On the same day, CNN reported that the scientist in 2006 had admitted faking his findings after questions of impropriety had emerged.
• In the same section (Timeline): In mid-December, co-author of Hwang's papers came forward, telling the media that Hwang had confessed to fabricating evidence for nine of the eleven cell lines.
• In the same section (Timeline): On December 29, 2005, the university determined that all 11 of Hwang's stem cell lines were fabricated. The university announced on January 10, 2006, that Hwang's 2004 and 2005 papers in Science were both fabricated. Following on the confirmation of scientific misconduct, on January 11, Science unconditionally retracted both of Hwang's papers.
• In the same section (Timeline): On January 20, 2006, Hwang maintained that two of his 11 forged stem cell lines had been maliciously switched for cells from regular, not cloned, embryos.
• Later (Controversies): Close scrutiny revealed that several of the photos of purportedly different cells were in fact photos of the same cell.
• Later (Controversies): Researchers raised questions about striking similarities between the DNA profiles of the cloned cells. Then collaborator Gerald Schatten asked Science to remove his name from the paper, stating as a reason that there were "allegations from someone involved with the experiments that certain elements of the report may be fabricated."
• Later (Controversies): The scandal took a dramatic turn on December 15, when Roh Sung-il, who had collaborated on the disputed paper, stated to media outlets that nine of those eleven lines had been faked; specifically, DNA tests illustrated that those nine lines shared identical DNA, implying that they had come from the same source. Roh stated that "Professor Hwang admitted to fabrication", and that he, Hwang, and another co-author had asked Science to withdraw the paper.
• Later (Controversies): He, however, acknowledged the falsifications of research data in the paper, attributing them to unrecoverable "artificial mistakes".
• Later (Controversies): He maintained that at least nine of the eleven stem cell lines were fakes and that Hwang was simply untrustworthy.
• Later (Controversies): Roh also told MBC television that Hwang had pressured a former scientist at his lab to fake data to make it look like there were 11 stem cell colonies. In a separate report, a former researcher told MBC that Hwang ordered him to fabricate photos to make it appear there were 11 separate colonies from only three.
• Later (Controversies): although he admitted to misconduct in his follow-up paper on patient-specific stem cells
• Later (Official probe and confirmation of fraud): Official probe and confirmation of fraud
• Later (Official probe and confirmation of fraud): On December 23, the panel announced its initial finding that Hwang had intentionally fabricated stem cell research results creating nine fake cell lines out of eleven, and added that the validity of two remaining cell lines is yet to be confirmed.
• Later (Official probe and confirmation of fraud): Hwang's claim of having used only 185 eggs to create stem cell lines was also denied by the panel, which indicated that more eggs may have been used in the research process.
• Later (Official probe and confirmation of fraud): The panel announced additional findings on December 29, and confirmed that no patient-matched embryonic stem cells existed, and that Hwang's team did not have the scientific data to prove any of the stem cells had ever been made.
• Later (Official probe and confirmation of fraud): The panel found out that, contrary to Hwang's claim of having used 185 eggs for his team's 2005 paper, at least 273 eggs were shown to have been used according to research records kept in Hwang's lab.
• Later (Official probe and confirmation of fraud): such a claim was incomprehensible when there were no data to prove any of the stem cells were ever made to begin with.
• Later (Official probe and confirmation of fraud): The panel, in conclusion, stated that Hwang's team intentionally fabricated the data in both the 2004 and the 2005 papers, as described by Myung Hee Chung (Head of Seoul National University's investigation) and that it was an act of "deception of the scientific community and the public at large".
• Later (Indictment of Hwang and collaborators): Also in 2005, Hwang received one billion won each in research funds from SK Group and the National Agricultural Cooperative Federation based on his fabricated stem cell research results.
• Later (Supporting lawmakers group): After most of Hwang's claims were confirmed to be false on January 10, 2006, some lawmakers revealed that Hwang had made several campaign donations to them and other lawmakers.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-1007 / 1962277 / #1, read from revision 1376690909 of Hwang Woo-suk.

The model's reasoning: This retells the targeted paper's patient-specific stem-cell and efficiency claims without explicitly identifying them as fabricated in this paragraph; any independent historical sourcing needs checking.

Where the model found it: His team, which cloned the first human embryo to use for research, said they had used the same technology to create batches of embryonic stem cells from nine patients. According to Hwang, the result was much more efficient than they had hoped. Hwang's integrity as a researcher was again put in doubt when it was revealed that PD Su-cheop had scheduled a follow-up report questioning his achievement published in Science in June 2005, which stated he had cloned 11 lines of embryonic stem cells. This caused furious backlash among many South Koreans, and the reaction only intensified when it was discovered that Kim Sun-Jong, one of Hwang's researchers from MizMedi, had been coerced by illegal means to testify against Hwang. As a result, the scheduled broadcast was canceled and the network made a public apology to the nation, everyone more or less operating under the assumption that the show was at fault and not Hwang. Yet, other news outlets began to question Hwang's claims.

Where the model says the article warns about this paper:

• In the opening paragraphs: However, he became infamous around November 2005 for fabricating a series of stem cell experiments that were published in high-profile journals in a case known as the Hwang affair.
• In the opening paragraphs: Shortly after this, data from his human cloning experiments was revealed to have been falsified.
• In the opening paragraphs: On May 12, 2006, Hwang was charged with embezzlement and bioethics law violations after it emerged much of his stem cell research had been faked.
• In the opening paragraphs: On the same day, CNN reported that the scientist in 2006 had admitted faking his findings after questions of impropriety had emerged.
• Earlier (Timeline): In mid-December, co-author of Hwang's papers came forward, telling the media that Hwang had confessed to fabricating evidence for nine of the eleven cell lines.
• Earlier (Timeline): On December 29, 2005, the university determined that all 11 of Hwang's stem cell lines were fabricated. The university announced on January 10, 2006, that Hwang's 2004 and 2005 papers in Science were both fabricated. Following on the confirmation of scientific misconduct, on January 11, Science unconditionally retracted both of Hwang's papers.
• Earlier (Timeline): On January 20, 2006, Hwang maintained that two of his 11 forged stem cell lines had been maliciously switched for cells from regular, not cloned, embryos.
• In the same section (Controversies): Close scrutiny revealed that several of the photos of purportedly different cells were in fact photos of the same cell.
• In the same section (Controversies): Researchers raised questions about striking similarities between the DNA profiles of the cloned cells. Then collaborator Gerald Schatten asked Science to remove his name from the paper, stating as a reason that there were "allegations from someone involved with the experiments that certain elements of the report may be fabricated."
• In the same section (Controversies): The scandal took a dramatic turn on December 15, when Roh Sung-il, who had collaborated on the disputed paper, stated to media outlets that nine of those eleven lines had been faked; specifically, DNA tests illustrated that those nine lines shared identical DNA, implying that they had come from the same source. Roh stated that "Professor Hwang admitted to fabrication", and that he, Hwang, and another co-author had asked Science to withdraw the paper.
• In the same section (Controversies): He, however, acknowledged the falsifications of research data in the paper, attributing them to unrecoverable "artificial mistakes".
• In the same section (Controversies): He maintained that at least nine of the eleven stem cell lines were fakes and that Hwang was simply untrustworthy.
• In the same section (Controversies): Roh also told MBC television that Hwang had pressured a former scientist at his lab to fake data to make it look like there were 11 stem cell colonies. In a separate report, a former researcher told MBC that Hwang ordered him to fabricate photos to make it appear there were 11 separate colonies from only three.
• In the same section (Controversies): although he admitted to misconduct in his follow-up paper on patient-specific stem cells
• In the same section (Official probe and confirmation of fraud): Official probe and confirmation of fraud
• In the same section (Official probe and confirmation of fraud): On December 23, the panel announced its initial finding that Hwang had intentionally fabricated stem cell research results creating nine fake cell lines out of eleven, and added that the validity of two remaining cell lines is yet to be confirmed.
• In the same section (Official probe and confirmation of fraud): Hwang's claim of having used only 185 eggs to create stem cell lines was also denied by the panel, which indicated that more eggs may have been used in the research process.
• In the same section (Official probe and confirmation of fraud): The panel announced additional findings on December 29, and confirmed that no patient-matched embryonic stem cells existed, and that Hwang's team did not have the scientific data to prove any of the stem cells had ever been made.
• In the same section (Official probe and confirmation of fraud): The panel found out that, contrary to Hwang's claim of having used 185 eggs for his team's 2005 paper, at least 273 eggs were shown to have been used according to research records kept in Hwang's lab.
• In the same section (Official probe and confirmation of fraud): such a claim was incomprehensible when there were no data to prove any of the stem cells were ever made to begin with.
• In the same section (Official probe and confirmation of fraud): The panel, in conclusion, stated that Hwang's team intentionally fabricated the data in both the 2004 and the 2005 papers, as described by Myung Hee Chung (Head of Seoul National University's investigation) and that it was an act of "deception of the scientific community and the public at large".
• In the same section (Indictment of Hwang and collaborators): Also in 2005, Hwang received one billion won each in research funds from SK Group and the National Agricultural Cooperative Federation based on his fabricated stem cell research results.
• Later (Supporting lawmakers group): After most of Hwang's claims were confirmed to be false on January 10, 2006, some lawmakers revealed that Hwang had made several campaign donations to them and other lawmakers.

The model's confidence: 0.97. Finding RW-1007 / 1962277 / #4, read from revision 1376690909 of Hwang Woo-suk.

The model's reasoning: This repeats Hwang's assertion that the patient-matched technology underlying the 2005 paper existed, without an explicit local refutation; an editor should check the sources for this historical statement.

Where the model found it: On December 23, 2005, Hwang apologized for "creating a shock and a disappointment" and announced that he was resigning his position as professor at the university. However, Hwang maintained that patient-matched stem cell technology remained in South Korea, and his countrymen would see it.

Where the model says the article warns about this paper:

• In the opening paragraphs: However, he became infamous around November 2005 for fabricating a series of stem cell experiments that were published in high-profile journals in a case known as the Hwang affair.
• In the opening paragraphs: Shortly after this, data from his human cloning experiments was revealed to have been falsified.
• In the opening paragraphs: On May 12, 2006, Hwang was charged with embezzlement and bioethics law violations after it emerged much of his stem cell research had been faked.
• In the opening paragraphs: On the same day, CNN reported that the scientist in 2006 had admitted faking his findings after questions of impropriety had emerged.
• Earlier (Timeline): In mid-December, co-author of Hwang's papers came forward, telling the media that Hwang had confessed to fabricating evidence for nine of the eleven cell lines.
• Earlier (Timeline): On December 29, 2005, the university determined that all 11 of Hwang's stem cell lines were fabricated. The university announced on January 10, 2006, that Hwang's 2004 and 2005 papers in Science were both fabricated. Following on the confirmation of scientific misconduct, on January 11, Science unconditionally retracted both of Hwang's papers.
• Earlier (Timeline): On January 20, 2006, Hwang maintained that two of his 11 forged stem cell lines had been maliciously switched for cells from regular, not cloned, embryos.
• In the same section (Controversies): Close scrutiny revealed that several of the photos of purportedly different cells were in fact photos of the same cell.
• In the same section (Controversies): Researchers raised questions about striking similarities between the DNA profiles of the cloned cells. Then collaborator Gerald Schatten asked Science to remove his name from the paper, stating as a reason that there were "allegations from someone involved with the experiments that certain elements of the report may be fabricated."
• In the same section (Controversies): The scandal took a dramatic turn on December 15, when Roh Sung-il, who had collaborated on the disputed paper, stated to media outlets that nine of those eleven lines had been faked; specifically, DNA tests illustrated that those nine lines shared identical DNA, implying that they had come from the same source. Roh stated that "Professor Hwang admitted to fabrication", and that he, Hwang, and another co-author had asked Science to withdraw the paper.
• In the same section (Controversies): He, however, acknowledged the falsifications of research data in the paper, attributing them to unrecoverable "artificial mistakes".
• In the same section (Controversies): He maintained that at least nine of the eleven stem cell lines were fakes and that Hwang was simply untrustworthy.
• In the same section (Controversies): Roh also told MBC television that Hwang had pressured a former scientist at his lab to fake data to make it look like there were 11 stem cell colonies. In a separate report, a former researcher told MBC that Hwang ordered him to fabricate photos to make it appear there were 11 separate colonies from only three.
• In the same section (Controversies): although he admitted to misconduct in his follow-up paper on patient-specific stem cells
• In the same section (Official probe and confirmation of fraud): Official probe and confirmation of fraud
• In the same section (Official probe and confirmation of fraud): On December 23, the panel announced its initial finding that Hwang had intentionally fabricated stem cell research results creating nine fake cell lines out of eleven, and added that the validity of two remaining cell lines is yet to be confirmed.
• In the same section (Official probe and confirmation of fraud): Hwang's claim of having used only 185 eggs to create stem cell lines was also denied by the panel, which indicated that more eggs may have been used in the research process.
• In the same section (Official probe and confirmation of fraud): The panel announced additional findings on December 29, and confirmed that no patient-matched embryonic stem cells existed, and that Hwang's team did not have the scientific data to prove any of the stem cells had ever been made.
• In the same section (Official probe and confirmation of fraud): The panel found out that, contrary to Hwang's claim of having used 185 eggs for his team's 2005 paper, at least 273 eggs were shown to have been used according to research records kept in Hwang's lab.
• In the same section (Official probe and confirmation of fraud): such a claim was incomprehensible when there were no data to prove any of the stem cells were ever made to begin with.
• In the same section (Official probe and confirmation of fraud): The panel, in conclusion, stated that Hwang's team intentionally fabricated the data in both the 2004 and the 2005 papers, as described by Myung Hee Chung (Head of Seoul National University's investigation) and that it was an act of "deception of the scientific community and the public at large".
• In the same section (Indictment of Hwang and collaborators): Also in 2005, Hwang received one billion won each in research funds from SK Group and the National Agricultural Cooperative Federation based on his fabricated stem cell research results.
• Later (Supporting lawmakers group): After most of Hwang's claims were confirmed to be false on January 10, 2006, some lawmakers revealed that Hwang had made several campaign donations to them and other lawmakers.

The model's confidence: 0.86. Finding RW-1007 / 1962277 / #5, read from revision 1376690909 of Hwang Woo-suk.

The model's reasoning: The title establishes relevance to hydrogen aviation, but without an abstract it is unclear whether the review supports characterizing hydrogen planes as speculative; the Economist article plausibly supports this claim.

Other citations on it:

• "Is the time now ripe for planes to run on hydrogen?". The Economist. ISSN 0013-0613. Retrieved 2024-02-17.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.85. Finding RW-72708 / 906156 / #1, read from revision 1378382087 of Hydrogen economy.

The model's reasoning: This states the review’s finding of reduced disease following widespread conjugate vaccination; the other review plausibly supports prevention of meningitis but not clearly sepsis.

Other citations on it:

• Patel M, Lee CK (January 2005). "Polysaccharide vaccines for preventing serogroup A meningococcal meningitis". The Cochrane Database of Systematic Reviews. 2012 (1) CD001093. doi:10.1002/14651858.CD001093.pub2. PMC 11852230. PMID 15674874. CD001093.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-33410 / 21053494 / #1, read from revision 1371333387 of Meningococcal vaccine.

The model's reasoning: This presents the review's lack of eligible randomized trials as evidence, while extending its scope beyond Chinese herbal products; the temporary removal supplies no finding-specific reason for rejecting the conclusion, but an unretracted source is needed.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-30807 / 64076718 / #3, read from revision 1378243759 of Idiopathic chronic fatigue.

The model's reasoning: This states the withdrawn review's vaccine finding as established effectiveness; the review concerned exacerbations of chronic bronchitis and expressed its conclusion tentatively.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-33526 / 302755 / #1, read from revision 1319165533 of Acute bronchitis.

The model's reasoning: This screening recommendation relies on the withdrawn guideline; because it was withdrawn as outdated, a current guideline is needed.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-39893 / 3202774 / #1, read from revision 1376082553 of Neural tube defect.

The model's reasoning: The comparative claims justify the screening recommendation using the outdated guideline and need verification against a current source.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.95. Finding RW-39893 / 3202774 / #2, read from revision 1376082553 of Neural tube defect.

The model's reasoning: This research recommendation may draw on the review's call for further studies of benzodiazepine receptor modulators, but its abstract does not specifically assess flumazenil.

The whole list item: Flumazenil has been found to stimulate the reversal of tolerance and the normalization of receptor function. However, further research is needed in the form of randomised trials to demonstrate its role in the treatment of benzodiazepine withdrawal. Flumazenil stimulates the up-regulation and reverses the uncoupling of benzodiazepine receptors to the GABAA receptor, thereby reversing tolerance and reducing withdrawal symptoms and relapse rates. Because of limited research and experience compared to the possible risks involved, the flumazenil detoxification method is controversial and can only be done as an inpatient procedure under medical supervision.

Other citations on it:

• Gerra, G.; Zaimovic, A.; Giusti, F.; Moi, G.; Brewer, C. (2002). "Intravenous flumazenil versus oxazepam tapering in the treatment of benzodiazepine withdrawal: A randomized, placebo-controlled study". Addiction Biology. 7 (4): 385–95. doi:10.1080/1355621021000005973. PMID 14578014. S2CID 21255719.
• Little, H.J. (1991). "The benzodiazepines: Anxiolytic and withdrawal effects". Neuropeptides. 19: 11–4. doi:10.1016/0143-4179(91)90077-V. PMID 1679209. S2CID 13734753.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.78. Finding RW-33038 / 11683572 / #1, read from revision 1373095815 of Benzodiazepine withdrawal syndrome.

The model's reasoning: The item presents the paper's apparent subject—attribution bias in consumer choice and its consequences—as fact; replacement is appropriate because the retraction concerns publisher duplication, not invalid findings.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.87. Finding RW-28634 / 523104 / #2, read from revision 1373839279 of Managerial economics.

The model's reasoning: These personality, morphology and foraging associations match the paper's stated topic; the dispersal-genetics citation does not plausibly support these specific associations.

The sentence the footnote ends: Smaller spiders tend to have a bold personality.

Other citations on it:

• Smith, D.R.; Su, Y.; Berger-Tal, R; Lubin, Y (2016), "Population genetic evidence for sex-specific dispersal in an inbred social spider", Ecology and Evolution, 6 (15): 5479–5490, Bibcode:2016EcoEv...6.5479S, doi:10.1002/ece3.2200, PMC 4984519, PMID 27551398

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.88. Finding RW-47382 / 59753398 / #5, read from revision 1374093906 of Stegodyphus dumicola.

The model's reasoning: This reports an observation about task participation and collective foraging, directly matching the paper's topic.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.87. Finding RW-47382 / 59753398 / #7, read from revision 1374093906 of Stegodyphus dumicola.

The model's reasoning: The paragraph reports colony-level foraging effects of group size, personality and morphology, matching the cited paper; duplication and salami slicing warrant seeking a non-retracted source rather than assuming the results false.

The sentence the footnote ends: If a group was made up of spiders that were morphologically diverse (varying prosoma width), they also mounted a faster attack.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.94. Finding RW-47382 / 59753398 / #8, read from revision 1374093906 of Stegodyphus dumicola.

The model's reasoning: This repeats the morphology–task-participation relationship central to the paper's title; an editor should check whether this paragraph has an independent source.

Where the model found it: Larger spiders are more likely to take part in cooperative prey capture. During capture, many females will engage with the prey at the same time, injecting their digestive juices into the prey through extra-oral digestion. These digestive enzymes break down the prey into a liquid for other spiders to share, or for females to regurgitate to spiderlings.

The model's confidence: 0.78. Finding RW-47382 / 59753398 / #9, read from revision 1374093906 of Stegodyphus dumicola.

The model's reasoning: This reports the retracted paper's detection finding; the subsequent negative results and contamination concerns do not explicitly identify it as retracted or definitively refuted in this paragraph.

Where the model says the article warns about this paper:

• In the opening paragraphs: has made discredited medical claims, such as that murine endogenous retroviruses are found in blood samples of most patients with myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS).
• In the opening paragraphs: However, following widespread criticism, the paper was retracted on December 22, 2011, by the journal Science.
• In the same section (XMRV and myalgic encephalomyelitis/chronic fatigue syndrome): Two of the original authors of this paper subsequently reanalyzed the samples used in the research and found that the samples were contaminated with XMRV plasmid DNA, leading them to publish a partial retraction of their original results. In December 2011, after a request by Silverman, the editors of Science retracted the paper in its entirety.

Other citations on it:

• Callaway, Ewen (March 14, 2011). "Virology: Fighting for a cause". Nature. 471 (7338): 282–85. Bibcode:2011Natur.471..282C. doi:10.1038/471282a. PMID 21412308.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.91. Finding RW-492 / 27419466 / #1, read from revision 1360216434 of Judy Mikovits.

The model's reasoning: Although the footnote marks publication information, the paragraph presents the paper's purported successful nuclear-transfer procedure as fact without a local retraction warning.

The sentence the footnote ends: The research, published in the 12 March 2004 issue of Science, was reported as "Evidence of a pluripotent human embryonic stem cell line derived from a cloned blastocyst."

Where the model says the article warns about this paper:

• In the opening paragraphs: The SNU immediately investigated the research work and found that both the 2004 and 2005 papers contained fabricated results. Hwang was compelled to resign from the university, and publicly confessed in January 2006 that the research papers were based on fabricated data. Science immediately retracted the two papers.
• Later (Documentary): In April 2023, Canadian Youtuber BobbyBroccoli released a two-part docu-series entitled "The Man who faked human cloning" and "How to catch a criminal cloner" on Youtube and on the streaming service Nebula.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.97. Finding RW-1004 / 73655972 / #1, read from revision 1358851083 of Hwang affair.

The model's reasoning: The paragraph quotes the fabricated study's conclusion and then endorses its claimed cloning achievement as fact.

The sentence the footnote ends: The report concluded: "This study shows the feasibility of generating human ES [embryonic stem] cells from a somatic cell isolated from a living person."

Where the model says the article warns about this paper:

• In the opening paragraphs: The SNU immediately investigated the research work and found that both the 2004 and 2005 papers contained fabricated results. Hwang was compelled to resign from the university, and publicly confessed in January 2006 that the research papers were based on fabricated data. Science immediately retracted the two papers.
• Later (Documentary): In April 2023, Canadian Youtuber BobbyBroccoli released a two-part docu-series entitled "The Man who faked human cloning" and "How to catch a criminal cloner" on Youtube and on the streaming service Nebula.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-1004 / 73655972 / #2, read from revision 1358851083 of Hwang affair.

The model's reasoning: This reports the study's interpretation of its purported cells; neither the caveat nor the section heading tells readers that the results were fabricated or retracted.

Where the model says the article warns about this paper:

• In the opening paragraphs: The SNU immediately investigated the research work and found that both the 2004 and 2005 papers contained fabricated results. Hwang was compelled to resign from the university, and publicly confessed in January 2006 that the research papers were based on fabricated data. Science immediately retracted the two papers.
• Later (Documentary): In April 2023, Canadian Youtuber BobbyBroccoli released a two-part docu-series entitled "The Man who faked human cloning" and "How to catch a criminal cloner" on Youtube and on the streaming service Nebula.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.86. Finding RW-1004 / 73655972 / #3, read from revision 1358851083 of Hwang affair.

The model's reasoning: This lead summary repeats the 2004 paper's central finding without explicitly refuting it in this paragraph; an editor should check whether its own sources adequately contextualize the reported claim.

Where the model found it: The Hwang affair, or Hwang scandal, or Hwanggate, is a case of scientific misconduct and ethical issues surrounding a South Korean biologist, Hwang Woo-suk, who claimed to have created the first human embryonic stem cells by cloning in 2004. Hwang and his research team at the Seoul National University reported in the journal Science that they successfully developed a somatic cell nuclear transfer method with which they made the stem cells. In 2005, they published again in Science the successful cloning of 11 person-specific stem cells using 185 human eggs. The research was hailed as "a ground-breaking paper" in science. Hwang was elevated as "the pride of Korea", "national hero" [of Korea], and a "supreme scientist", to international praise and fame. Recognitions and honours immediately followed, including South Korea's Presidential Award in Science and Technology, and Time magazine listing him among the "People Who Mattered 2004" and the most influential people "The 2004 Time 100".

Where the model says the article warns about this paper:

• In the opening paragraphs: The SNU immediately investigated the research work and found that both the 2004 and 2005 papers contained fabricated results. Hwang was compelled to resign from the university, and publicly confessed in January 2006 that the research papers were based on fabricated data. Science immediately retracted the two papers.
• Later (Documentary): In April 2023, Canadian Youtuber BobbyBroccoli released a two-part docu-series entitled "The Man who faked human cloning" and "How to catch a criminal cloner" on Youtube and on the streaming service Nebula.

The model's confidence: 0.93. Finding RW-1004 / 73655972 / #4, read from revision 1358851083 of Hwang affair.

The model's reasoning: This adjacent paragraph repeats the paper's claimed genetic identity and differentiation results and draws a therapeutic conclusion; its own footnotes need checking for independent support and historical framing.

Where the model found it: Hwang publicly reported the research at the annual meeting of the American Association for the Advancement of Science (AAAS) in Seattle on 16 February 2004. He specified that they used 242 eggs from 16 unpaid volunteers, creating about 100 cells from which 30 embryos were developed. Since the embryos had adult DNA, the resulting stem cells became clones of the adult somatic cells. From the embryos, the stem cells were collected and grafted into mice in which they could grow into various body parts including muscle, bone, cartilage and connective tissues. The method ensured that immune rejection would be avoided so that it could be used for the treatment of genetic disorders. As Hwang explained: "Our approach opens the door for the use of these specially developed cells in transplantation medicine."

Where the model says the article warns about this paper:

• In the opening paragraphs: The SNU immediately investigated the research work and found that both the 2004 and 2005 papers contained fabricated results. Hwang was compelled to resign from the university, and publicly confessed in January 2006 that the research papers were based on fabricated data. Science immediately retracted the two papers.
• Later (Documentary): In April 2023, Canadian Youtuber BobbyBroccoli released a two-part docu-series entitled "The Man who faked human cloning" and "How to catch a criminal cloner" on Youtube and on the streaming service Nebula.

The model's confidence: 0.98. Finding RW-1004 / 73655972 / #5, read from revision 1358851083 of Hwang affair.

The model's reasoning: These sentences report the fabricated cloning and patient-matched stem-cell results without a local retraction or refutation warning.

The sentence the footnote ends: Hwang's team reported another successful cloning of human cells in the 17 June 2005 issue of Science, in this case, embryonic stem cells derived from skin cells.

Where the model says the article warns about this paper:

• In the opening paragraphs: The SNU immediately investigated the research work and found that both the 2004 and 2005 papers contained fabricated results. Hwang was compelled to resign from the university, and publicly confessed in January 2006 that the research papers were based on fabricated data. Science immediately retracted the two papers.
• Later (Documentary): "The Man who faked human cloning"

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-1007 / 73655972 / #1, read from revision 1358851083 of Hwang affair.

The model's reasoning: The experimental counts and donation claims rest on the retracted report; this historical account does not locally explain its retraction.

The sentence the footnote ends: The report explicitly stated that: "Patients voluntarily donated oocytes and somatic cells for therapeutic cloning research and relevant applications but not for reproductive cloning ... no financial reimbursement in any form was paid."

Where the model says the article warns about this paper:

• In the opening paragraphs: The SNU immediately investigated the research work and found that both the 2004 and 2005 papers contained fabricated results. Hwang was compelled to resign from the university, and publicly confessed in January 2006 that the research papers were based on fabricated data. Science immediately retracted the two papers.
• Later (Documentary): "The Man who faked human cloning"

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.95. Finding RW-1007 / 73655972 / #2, read from revision 1358851083 of Hwang affair.

The model's reasoning: The lead repeats the retracted paper's distinctive stem-cell count and egg count without citing it here; an editor should check whether any local source independently contextualizes these claims.

Where the model found it: The Hwang affair, or Hwang scandal, or Hwanggate, is a case of scientific misconduct and ethical issues surrounding a South Korean biologist, Hwang Woo-suk, who claimed to have created the first human embryonic stem cells by cloning in 2004. Hwang and his research team at the Seoul National University reported in the journal Science that they successfully developed a somatic cell nuclear transfer method with which they made the stem cells. In 2005, they published again in Science the successful cloning of 11 person-specific stem cells using 185 human eggs. The research was hailed as "a ground-breaking paper" in science. Hwang was elevated as "the pride of Korea", "national hero" [of Korea], and a "supreme scientist", to international praise and fame. Recognitions and honours immediately followed, including South Korea's Presidential Award in Science and Technology, and Time magazine listing him among the "People Who Mattered 2004" and the most influential people "The 2004 Time 100".

Where the model says the article warns about this paper:

• In the opening paragraphs: The SNU immediately investigated the research work and found that both the 2004 and 2005 papers contained fabricated results. Hwang was compelled to resign from the university, and publicly confessed in January 2006 that the research papers were based on fabricated data. Science immediately retracted the two papers.
• Later (Documentary): "The Man who faked human cloning"

The model's confidence: 0.98. Finding RW-1007 / 73655972 / #3, read from revision 1358851083 of Hwang affair.

The model's reasoning: The claim uses the withdrawn review as evidence of disease prevention, although it found reduced gingivitis and no trials reporting caries outcomes; the other systematic review plausibly supports the caries component.

Other citations on it:

• Hujoel, P.P.; Cunha-Cruz, J.; Banting, D.W.; Loesche, W.J. (2006). "Dental Flossing and Interproximal Caries: a Systematic Review". Journal of Dental Research. 85 (4): 298–305. doi:10.1177/154405910608500404. ISSN 0022-0345. PMID 16567548. S2CID 53789465.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-30653 / 63116845 / #1, read from revision 1374331478 of Interdental cleaning.

The model's reasoning: This repeats the flossing-related plaque and disease-prevention claim supported by the withdrawn review in the lead, extending beyond its findings; an editor must check whether this paragraph has independent supporting sources.

Where the model found it: Floss is one of the most commonly used interdental cleaners. It is traditionally made of waxed nylon wrapped up in a plastic box. Since dental floss is able to remove some inter-proximal plaque, frequent regular dental flossing will reduce inter-proximal caries and periodontal disease risks. Results of a high-level of evidence meta-analysis recently found that floss may not be the most effective method for interdental cleaning, contrary to common belief at the time. Especially for individuals lacking dexterity or compliance, waterjet irrigators and interdental brushes were found to be significantly more effective than flossing.

The model's confidence: 0.85. Finding RW-30653 / 63116845 / #2, read from revision 1374331478 of Interdental cleaning.

The model's reasoning: The statement about lack of evidence rests on the review's conclusion of insufficient comparative evidence; the review was withdrawn as out of date and replaced.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-30825 / 18427821 / #1, read from revision 1322165958 of Facial trauma.

The model's reasoning: The assessment of weak evidence for pediatric RSV treatment reflects this review's conclusion that trials lacked power; the paper was withdrawn as out of date.

The sentence the footnote ends: It is noted by the United States Army Medical Research Institute of Infectious Diseases (USAMRIID) that "Ribavirin has poor in vitro and in vivo activity against the filoviruses (Ebola and Marburg) and the flaviviruses (dengue, yellow fever, Omsk hemorrhagic fever, and Kyasanur forest disease)" The aerosol form has been used in the past to treat respiratory syncytial virus-related diseases in children, although the evidence to support this is rather weak.

Other citations on it:

• Goeijenbier M, van Kampen JJ, Reusken CB, Koopmans MP, van Gorp EC (November 2014). "Ebola virus disease: a review on epidemiology, symptoms, treatment and pathogenesis". The Netherlands Journal of Medicine. 72 (9): 442–448. PMID 25387613. Archived from the original on 2014-11-29.
• Medical Management of Biological Casualties Handbook. United States Government Printing Office. 2011. p. 115. ISBN 978-0-16-090015-0.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-33569 / 71581 / #1, read from revision 1359690341 of Ribavirin.

The model's reasoning: The sentence repeats the review's conclusions about surface roughness, peri-implantitis and long-term success; the review was withdrawn as out of date.

The whole list item: Root form implants; the most common type of implant indicated for all uses. Within the root form type of implant, there are roughly 18 variants, all made of titanium but with different shapes and surface textures. There is limited evidence showing that implants with relatively smooth surfaces are less prone to peri-implantitis than implants with rougher surfaces and no evidence showing that any particular type of dental implant has superior long-term success.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-30209 / 59781843 / #1, read from revision 1377614465 of History of dental treatments.

The model's reasoning: This specific structural communication mechanism appears to rest on the paper's proposed coupling between nucleotide-binding and transmembrane domains; an editor must check whether this paragraph has independent support.

Where the model found it: MsbA (3.A.1.106.1) is a multi-drug resistant (MDR) ABC transporter and possibly a lipid flippase. It is an ATPase that transports lipid A, the hydrophobic moiety of lipopolysaccharide (LPS), a glucosamine-based saccharolipid that makes up the outer monolayer of the outer membranes of most gram-negative bacteria. Lipid A is an endotoxin and so loss of MsbA from the cell membrane or mutations that disrupt transport results in the accumulation of lipid A in the inner cell membrane resulting to cell death. It is a close bacterial homolog of P-glycoprotein (Pgp) by protein sequence homology and has overlapping substrate specificities with the MDR-ABC transporter LmrA from Lactococcus lactis. MsbA from E. coli is 36% identical to the NH2-terminal half of human MDR1, suggesting a common mechanism for transport of amphiphatic and hydrophobic substrates. The MsbA gene encodes a half transporter that contains a transmembrane domain (TMD) fused with a nucleotide-binding domain (NBD). It is assembled as a homodimer with a total molecular mass of 129.2 kD. MsbA contains 6 TMDs on the periplasmic side, an NBD located on the cytoplasmic side of the cell membrane, and an intracellular domain (ICD), bridging the TMD and NBD. This conserved helix extending from the TMD segments into or near the active site of the NBD is largely responsible for crosstalk between TMD and NBD. In particular, ICD1 serves as a conserved pivot about which the NBD can rotate, therefore allowing the NBD to disassociate and dimerize during ATP binding and hydrolysis.

Where the model says the article warns about this paper:

• In the same section (MsbA): Previously published (and now retracted) X-ray structures of MsbA were inconsistent with the bacterial homolog Sav1866. The structures were reexamined and found to have an error in the assignment of the hand resulting to incorrect models of MsbA.

The model's confidence: 0.76. Finding RW-1000 / 1551873 / #2, read from revision 1374757828 of ABC transporter.

The model's reasoning: This repeats the paper's distinctive proposed flip-flop mechanism, including sequestered sugar groups and lipid tails dragged through the bilayer, without a local retraction warning; an editor must check any sources attached to this paragraph.

Where the model found it: ABC exporters have a transport mechanism that is consistent with both the alternating-access model and ATP-switch model. In the apo states of exporters, the conformation is inward-facing and the TMDs and NBDs are relatively far apart to accommodate amphiphilic or hydrophobic substrates. For MsbA, in particular, the size of the chamber is large enough to accommodate the sugar groups from lipopolysaccharides (LPS). As has been suggested by several groups, binding of substrate initiates the transport cycle. The "power stroke", that is, ATP binding that induces NBD dimerization and formation of the ATP sandwich, drives the conformational changes in the TMDs. In MsbA, the sugar head groups are sequestered within the chamber during the "power stroke". The cavity is lined with charged and polar residues that are likely solvated creating an energetically unfavorable environment for hydrophobic substrates and energetically favorable for polar moieties in amphiphilic compounds or sugar groups from LPS. Since the lipid cannot be stable for a long time in the chamber environment, lipid A and other hydrophobic molecules may "flip" into an energetically more favorable position within the outer membrane leaflet. The "flipping" may also be driven by the rigid-body shearing of the TMDs while the hydrophobic tails of the LPS are dragged through the lipid bilayer. Repacking of the helices switches the conformation into an outward-facing state. ATP hydrolysis may widen the periplasmic opening and push the substrate towards the outer leaflet of the lipid bilayer. Hydrolysis of the second ATP molecule and release of Pi separates the NBDs followed by restoration of the resting state, opening the chamber towards the cytoplasm for another cycle.

Where the model says the article warns about this paper:

• In the same section (MsbA): Previously published (and now retracted) X-ray structures of MsbA were inconsistent with the bacterial homolog Sav1866. The structures were reexamined and found to have an error in the assignment of the hand resulting to incorrect models of MsbA.

The model's confidence: 0.98. Finding RW-1000 / 1551873 / #3, read from revision 1374757828 of ABC transporter.

The model's reasoning: This repeats the paper's homodimer and six-helix structural description; an editor must check whether this paragraph independently cites corrected structures supporting these details.

Where the model found it: MsbA (3.A.1.106.1) is a multi-drug resistant (MDR) ABC transporter and possibly a lipid flippase. It is an ATPase that transports lipid A, the hydrophobic moiety of lipopolysaccharide (LPS), a glucosamine-based saccharolipid that makes up the outer monolayer of the outer membranes of most gram-negative bacteria. Lipid A is an endotoxin and so loss of MsbA from the cell membrane or mutations that disrupt transport results in the accumulation of lipid A in the inner cell membrane resulting to cell death. It is a close bacterial homolog of P-glycoprotein (Pgp) by protein sequence homology and has overlapping substrate specificities with the MDR-ABC transporter LmrA from Lactococcus lactis. MsbA from E. coli is 36% identical to the NH2-terminal half of human MDR1, suggesting a common mechanism for transport of amphiphatic and hydrophobic substrates. The MsbA gene encodes a half transporter that contains a transmembrane domain (TMD) fused with a nucleotide-binding domain (NBD). It is assembled as a homodimer with a total molecular mass of 129.2 kD. MsbA contains 6 TMDs on the periplasmic side, an NBD located on the cytoplasmic side of the cell membrane, and an intracellular domain (ICD), bridging the TMD and NBD. This conserved helix extending from the TMD segments into or near the active site of the NBD is largely responsible for crosstalk between TMD and NBD. In particular, ICD1 serves as a conserved pivot about which the NBD can rotate, therefore allowing the NBD to disassociate and dimerize during ATP binding and hydrolysis.

Where the model says the article warns about this paper:

• In the same section (MsbA): Previously published (and now retracted) X-ray structures of MsbA were inconsistent with the bacterial homolog Sav1866. The structures were reexamined and found to have an error in the assignment of the hand resulting to incorrect models of MsbA. Recently, the errors have been rectified and new structures have been reported.

The model's confidence: 0.83. Finding RW-1002 / 1551873 / #2, read from revision 1374757828 of ABC transporter.

The model's reasoning: This echoes the paper's proposed transport mechanism based on charged residues lining a central chamber; an editor must check whether the paragraph's own sources establish this interpretation from corrected structures.

Where the model found it: ABC exporters have a transport mechanism that is consistent with both the alternating-access model and ATP-switch model. In the apo states of exporters, the conformation is inward-facing and the TMDs and NBDs are relatively far apart to accommodate amphiphilic or hydrophobic substrates. For MsbA, in particular, the size of the chamber is large enough to accommodate the sugar groups from lipopolysaccharides (LPS). As has been suggested by several groups, binding of substrate initiates the transport cycle. The "power stroke", that is, ATP binding that induces NBD dimerization and formation of the ATP sandwich, drives the conformational changes in the TMDs. In MsbA, the sugar head groups are sequestered within the chamber during the "power stroke". The cavity is lined with charged and polar residues that are likely solvated creating an energetically unfavorable environment for hydrophobic substrates and energetically favorable for polar moieties in amphiphilic compounds or sugar groups from LPS. Since the lipid cannot be stable for a long time in the chamber environment, lipid A and other hydrophobic molecules may "flip" into an energetically more favorable position within the outer membrane leaflet. The "flipping" may also be driven by the rigid-body shearing of the TMDs while the hydrophobic tails of the LPS are dragged through the lipid bilayer. Repacking of the helices switches the conformation into an outward-facing state. ATP hydrolysis may widen the periplasmic opening and push the substrate towards the outer leaflet of the lipid bilayer. Hydrolysis of the second ATP molecule and release of Pi separates the NBDs followed by restoration of the resting state, opening the chamber towards the cytoplasm for another cycle.

Where the model says the article warns about this paper:

• In the same section (MsbA): Previously published (and now retracted) X-ray structures of MsbA were inconsistent with the bacterial homolog Sav1866. The structures were reexamined and found to have an error in the assignment of the hand resulting to incorrect models of MsbA. Recently, the errors have been rectified and new structures have been reported.

The model's confidence: 0.78. Finding RW-1002 / 1551873 / #3, read from revision 1374757828 of ABC transporter.

The model's reasoning: The reported assertion repeats the disputed STAP finding without a local retraction or refutation warning; an editor must check whether this paragraph's own source adequately contextualizes the quotation.

Where the model found it: In 2016, Obokata's book Ano hi (あの日- That Day) was published by Kodansha. In her account of the controversy, Obokata relates her association with Wakayama, writing that "crucial parts of the STAP experiments were handled only by Wakayama", that she received the STAP cells from Wakayama, and that Wakayama "changed his accounts of how the STAP cells were produced." Obokata later wrote "I feel a strong sense of responsibility for the STAP papers [...], "I never wrote those papers to deceive anyone," and "STAP was real."

Where the model says the article warns about this paper:

• In the opening paragraphs: Attempts to replicate Obokata's STAP cell results failed.
• In the same section (STAP cell controversy): Within days of publication of the Nature articles, "disturbing allegations emerged [...] images looked doctored, and chunks of [...] text were lifted from other papers."
• In the same section (STAP cell controversy): In manipulating the image data of two different gels and using data from two different experiments, Dr. Obokata acted in a manner that can by no means be permitted.
• In the same section (STAP cell controversy): In June 2014, Obokata agreed to retract both papers.
• In the same section (STAP cell controversy): Near the time of retraction, "genetic analysis showed that the STAP cells didn’t match the mice from which they supposedly came."
• In the same section (STAP cell controversy): "the obvious, and rather depressing, explanation is that her so-called STAP cells were just regular embryonic stem cells that someone had taken from a freezer and relabelled."
• In the same section (STAP cell controversy): Those efforts failed to replicate the results originally reported.
• In the same section (STAP cell controversy): In a February 2015 article, The Guardian reported that Obokata was guilty of "unbelievable carelessness", having "manipulated images and plagiarised text."

The model's confidence: 0.88. Finding RW-2086 / 41787575 / #4, read from revision 1372543632 of Haruko Obokata.

The model's reasoning: This states the review's conclusion about the lack of treatment evidence as current; the accompanying systematic review plausibly supports it but is also old, so an up-to-date source is needed.

Other citations on it:

• Chalmers RJ, O'Sullivan T, Owen CM, Griffiths CE (December 2001). "A systematic review of treatments for guttate psoriasis". The British Journal of Dermatology. 145 (6): 891–4. doi:10.1046/j.1365-2133.2001.04505.x. PMID 11899141. S2CID 27381477.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.98. Finding RW-30656 / 21099851 / #1, read from revision 1328545168 of Guttate psoriasis.

The model's reasoning: This states the retracted paper's specific Ax21-recognition finding; the later paper's title supports XA21 recognition of a sulfated protein but does not establish that it is Ax21.

The sentence the footnote ends: The PRR class of R genes includes the rice XA21 resistance gene that recognizes the ax21 peptide and the Arabidopsis FLS2 peptide that recognizes the flg22 peptide from flagellin.

Other citations on it:

• Pruitt RN, Schwessinger B, Joe A, Thomas N, Liu F, Albert M, et al. (July 2015). "The rice immune receptor XA21 recognizes a tyrosine-sulfated protein from a Gram-negative bacterium". Science Advances. 1 (6) e1500245. Bibcode:2015SciA....1E0245P. doi:10.1126/sciadv.1500245. PMC 4646787. PMID 26601222.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-498 / 8938711 / #1, read from revision 1342480562 of Gene-for-gene relationship.

The model's reasoning: The XA21 portion appears to repeat the retracted peptide-binding finding cited earlier; an editor should check whether this paragraph has an independent source establishing a different XA21 ligand.

Where the model found it: In only some cases is there direct interaction between the R gene product and the Avr gene product. For example, both FLS2 and XA21 interact with the microbial peptides. In contrast, for the NBS-LRR class of R genes, direct interaction has not been shown for most of the R/avr pairs. This lack of evidence for a direct interaction led to the formation of the guard hypothesis for the NBS-LRR class of R genes.

The model's confidence: 0.85. Finding RW-498 / 8938711 / #3, read from revision 1342480562 of Gene-for-gene relationship.

The model's reasoning: The row presents TREEFINDER's capabilities as fact, relying on the paper describing the software; the stated retraction reasons do not directly undermine those capabilities, but an unretracted source is needed.

The whole table row: Treefinder Fast ML tree reconstruction, bootstrap analysis, model selection, hypothesis testing, tree calibration, tree manipulation and visualization, computation of sitewise rates, sequence simulation, many models of evolution (DNA, protein, rRNA, mixed protein, user-definable), GUI and scripting language Maximum likelihood, distances, and others Jobb G, von Haeseler A, Strimmer K

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.97. Finding RW-521 / 6022892 / #1, read from revision 1377217948 of List of phylogenetics software.

The model's reasoning: Under the interaction-list introduction, this item asserts a SMARCB1–BAZ1B (WSTF) interaction supported only by the retracted WINAC study.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-216 / 14304380 / #1, read from revision 1300576965 of SMARCB1.

The model's reasoning: The paragraph reports the paper's cervical-tissue and HeLa expression findings as true, although this footnote ends on the preceding methods sentence.

The sentence the footnote ends: In one study studying the role of Cdc42 in cervical cancer, immunohistochemistry was used to detect Cdc42 expression in three types of tissues: normal cervical tissues, cervical intraepithelial neoplasia (CIN) I or below, CIN II or above, and cervical cancer tissues.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-68786 / 12044406 / #1, read from revision 1374944749 of CDC42.

The model's reasoning: This directly reports the retracted paper's migration result without qualification; concerns about unreliable and duplicated data undermine its sole support.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-68786 / 12044406 / #2, read from revision 1374944749 of CDC42.

The model's reasoning: These sentences repeat the paper's filopodia finding and migration conclusion without disclosing its retraction, and no other support is supplied.

The sentence the footnote ends: Cdc42 overexpression significantly improved the ability of cervical cancer cells to migrate, possibly due to improved pseudopodia formation.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-68786 / 12044406 / #3, read from revision 1374944749 of CDC42.

The model's reasoning: The sentence presents the review’s pain-reduction conclusion as valid; the supplied abstract does not establish the additional safety claim, and the reason for removal is unspecified.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-32245 / 647283 / #1, read from revision 1378261420 of Lidocaine/prilocaine.

The model's reasoning: This retells the paper's specific Aβ*56 findings without a local retraction or refutation warning; the Star Tribune attribution reports those same findings rather than independent validation, and an editor should check the paragraph's sources.

Where the model found it: Lesné is the lead author, with Ashe and others as co-authors, of a controversial 2006 report published in Nature. The paper, "A specific amyloid-β protein assembly in the brain impairs memory", suggested the Aβ*56 oligomer (known as amyloid beta star 56 and Abeta*56) as a cause of Alzheimer's disease. The study proposed that Aβ*56 was responsible for the memory deficits that occur in Alzheimer's. According to the Star Tribune, images from the study showed the Aβ*56 protein growing as mice began to decline and age with dementia.

Where the model says the article warns about this paper:

• In the opening paragraphs: Lesné's work in the 2006 publication and others has been investigated since June 2022 on charges that he manipulated images to inflate the role of Aβ*56 in Alzheimer's.
• In the opening paragraphs: Retracted in 2024, the paper was foundational in the hypothesis that one specific toxic oligomer of the amyloid beta protein, known as Aβ*56, caused memory impairment in Alzheimer's, aligned with the prevailing amyloid hypothesis.
• In the opening paragraphs: Karen Ashe, the senior author of the 2006 Nature paper – and all authors except Lesné – agreed to retract the paper which, according to Retraction Watch, makes it the second most highly cited paper ever retracted.
• In the same section (2022 investigation): A July 2022 publication in Science authored by Charles Piller questioned the authenticity of Western blot images used in Lesné's research; Piller's report alleges, based on a six-month investigation, that images may have been doctored to emphasize the role of Aβ*56 in Alzheimer's. Matthew Schrag, a Vanderbilt University neuroscientist, raised concerns in a whistleblower report that images were falsified, casting doubt on the association between the Aβ*56 protein and dementia symptoms.
• In the same section (2022 investigation): The editors of Nature responded with a July 14, 2022 note stating that they were aware of and investigating the concerns and that "readers are advised to use caution when using results reported therein".
• In the same section (2022 investigation): USA Today reported that Ashe had started the process to retract the Nature paper, but that had not been done as of July 29, 2022 because six of the eight co-authors would not sign off on the request.
• In the same section (2024 retraction): ### 2024 retraction
• In the same section (2024 retraction): In May 2024, Ashe announced that the 2006 publication would be retracted, although she stated that the manipulated images did not affect the study's conclusions; according to Retraction Watch, this makes it the second most highly cited paper ever retracted.
• In the same section (2024 retraction): Nature published retraction of the article on June 24, 2024, noting that the edited images showed "excessive manipulation, including splicing, duplication and the use of an eraser tool".
• In the same section (Impact on Alzheimer's research): Olson wrote that the allegations "wouldn't scuttle the entire theory of amyloid proteins", but "Selkoe told Science that they undercut the existence of the Aβ*56 protein that is central to Lesné's research."
• In the same section (Impact on Alzheimer's research): Selkoe, who Science describes as "a leading advocate of the amyloid and toxic oligomer hypotheses", opined that the "broader amyloid hypothesis remains viable" and referred to the alleged image manipulation as "what really looks like a very egregious example of malfeasance that's squarely in the Aβ oligomer field".
• Later (Legacy): In a February 13, 2025 book review of Charles Piller's Doctored: Fraud, Arrogance, and Tragedy in the Quest to Cure Alzheimer's, Carl Elliott wrote that "shocking was that some of Lesné's work had come from the lab of Karen Ashe, one of the most acclaimed neuroscientists in the field", noting that the University of Minnesota had the "dubious distinction of producing two of the five most highly cited scientific papers ever to be retracted" according to Retraction Watch.

The model's confidence: 0.98. Finding RW-54364 / 71365995 / #3, read from revision 1375081153 of Sylvain Lesné.

The model's reasoning: The caption presents the paper's regional economic-damage projections and statistical comparisons as valid results; the retraction concerns data, methods and analytical errors.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-72836 / 56868542 / #1, read from revision 1378412020 of Climate change in Asia.

The model's reasoning: This attributes the retracted paper's binding and opioid-tolerance findings without disclosing its retraction in this paragraph; the other sources concern Alzheimer's disease rather than independent confirmation of this finding.

The sentence the footnote ends: Burns and Wang reported in 2008 that FLNA contains the high-affinity binding site of naloxone and naltrexone in preventing opioid tolerance and dependence, and in 2020 that by disrupting that simufilam reduces the ultra-tight binding of amyloid beta 42 to the alpha-7 nicotinic receptor.

Where the model says the article warns about this paper:

• Earlier (History): Both of these papers were retracted by PLOS One in 2022.
• Later (Allegations of research irregularities): Two papers were retracted by journals and expressions of concern were issued for other papers.
• Later (Allegations of research irregularities): Two papers unrelated to Alzheimer's disease that reported FLNA binding by certain opioid antagonists and FLNA's role in opioid tolerance and dependence were retracted for "similarities in background pixels" in western blot images without evidence of data manipulation.

Other citations on it:

• Wang HY, Pei Z, Lee KC, Lopez-Brignoni E, Nikolov B, Crowley CA, et al. (2020). "PTI-125 Reduces Biomarkers of Alzheimer's Disease in Patients". The Journal of Prevention of Alzheimer's Disease. 7 (4): 256–264. doi:10.14283/jpad.2020.6. PMID 32920628. S2CID 211039039.
• Wang HY, Bakshi K, Frankfurt M, Stucky A, Goberdhan M, Shah SM, et al. (July 2012). "Reducing amyloid-related Alzheimer's disease pathogenesis by a small molecule targeting filamin A". The Journal of Neuroscience. 32 (29): 9773–9784. doi:10.1523/JNEUROSCI.0354-12.2012. PMC 6621293. PMID 22815492. (This paper currently has an expression of concern, see doi:10.1523/JNEUROSCI.2306-21.2021, PMID 34921050, Retraction Watch)

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-36319 / 70579655 / #3, read from revision 1362085996 of Simufilam.

The model's reasoning: The quotation presents the study's specialty differences and medical-school predictors as valid findings without disclosing the retraction; because this was a retract-and-replace case, the replacement needs checking.

The sentence the footnote ends: "Our data show wide variability in the prevalence of burnout by clinical specialty, and that anxiety, social support and empathy during medical school relate to the risk of burnout during residency," says Liselotte Dyrbye, M.D., a Mayo Clinic researcher and first author of the article "Association of Clinical Specialty With Symptoms of Burnout and Career Choice Regret Among US Resident Physicians"

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-20190 / 60275503 / #1, read from revision 1376351408 of Physician burnout.

The model's reasoning: The sentence presents the paper’s functional mapping of PEDF peptides as established findings, with no other citation supporting it.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-73382 / 14723340 / #1, read from revision 1378026922 of PEDF.

The model's reasoning: The characterization of residues 24-57 as an antiangiogenic region repeats the retracted paper’s peptide mapping; the receptor interaction may have an independent source that an editor should check.

Where the model found it: Secreted PEDF binds a receptor on the cell surface termed PEDF-R. PEDF-R has phospholipase A2 activity which liberates fatty acids from glycerolipids. PEDF enhances gamma-secretase activity, leading to the cleavage of the VEGF receptor 1 (VEGFR-1) transmembrane domain. This action interferes with VEGF signaling thereby inhibiting angiogenesis. Laminin receptor is also a target for PEDF, and the interaction occurs between residues 24-57 of PEDF, a region known to regulate antiangiogenic function.

The model's confidence: 0.88. Finding RW-73382 / 14723340 / #2, read from revision 1378026922 of PEDF.

The model's reasoning: This repeats the paper’s specific finding that the 34-mer inhibits angiogenesis; an editor should check whether this paragraph has independent supporting sources.

Where the model found it: PEDF has a variety of functions including antiangiogenic, antitumorigenic, and neurotrophic properties. Endothelial cell migration is inhibited by PEDF. PEDF suppresses retinal neovascularization and endothelial cell proliferation. The antiangiogenic residues 24-57 were shown to be sufficient at inhibiting angiogenesis. PEDF is also responsible for apoptosis of endothelial cells either through the p38 MAPK pathway or through the FAS/FASL pathway Antiangiogenic function is also conferred by PEDF through inhibition of both VEGFR-1 and VEGFR-2.

The model's confidence: 0.99. Finding RW-73382 / 14723340 / #3, read from revision 1378026922 of PEDF.

The model's reasoning: The halitosis-management claim draws on the review's tongue-cleaning results; the periodontology textbook plausibly supports it independently, and the review was retracted and replaced.

The whole list item: Tongue coating - food debris, desquamated epithelial cells and bacteria often form a visible tongue coating. This coating has been identified as a major contributing factor in bad breath (halitosis), which can be managed by brushing the tongue gently with a toothbrush or using special oral hygiene instruments such as tongue scrapers or mouth brushes.

Other citations on it:

• Newman MG, Takei HH, Klokkevold PR, Carranza FA, eds. (2012). Carranza's clinical periodontology (11th ed.). St. Louis, Mo.: Elsevier/Saunders. pp. 84–96. ISBN 978-1-4377-0416-7.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-31398 / 17844308 / #1, read from revision 1314808670 of Tongue disease.

The model's reasoning: The citation supports this structural claim from the review; text plagiarism does not itself establish that the claim is false, but an independent source is needed.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-7815 / 1041856 / #1, read from revision 1374052512 of Porin (protein).

The model's reasoning: The sentence presents the review's account of porin selectivity as fact; the plagiarism retraction calls for independent sourcing.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-7815 / 1041856 / #2, read from revision 1374052512 of Porin (protein).

The model's reasoning: This functional claim relies on the cited review's account of outer-membrane transport and needs a non-retracted source.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-7815 / 1041856 / #3, read from revision 1374052512 of Porin (protein).

The model's reasoning: The footnote plausibly covers both the preceding immune-signaling claim and the treatment claim, matching the review's focus on immune responses and therapeutic targets.

The sentence the footnote ends: Therapeutic treatments, like vaccinations and antibiotics, are used to supplement this immune response.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.94. Finding RW-7815 / 1041856 / #4, read from revision 1374052512 of Porin (protein).

The model's reasoning: The sentence uses the review as evidence for a porin-mediated resistance mechanism; plagiarism alone does not invalidate the mechanism, but independent sourcing is needed.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-7815 / 1041856 / #5, read from revision 1374052512 of Porin (protein).

The model's reasoning: The closing citation plausibly supports both the transport mechanism and its claimed public-health consequences; each requires independent verification and sourcing.

The sentence the footnote ends: These changes have contributed to the global emergence of antibiotic resistance, and an increase in mortality rates from infections.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.96. Finding RW-7815 / 1041856 / #6, read from revision 1374052512 of Porin (protein).

The model's reasoning: This repeats the amino-acid basis of selectivity explicitly supported by the retracted review under Cellular roles; an editor must check whether this separate paragraph has independent sourcing.

Where the model found it: Porins are composed of beta sheets (β sheets) made up of beta strands (β strands) which are linked together by beta turns (β turns) on the cytoplasmic side and long loops of amino acids on the other. The β strands lie in an antiparallel fashion and form a cylindrical tube, called a beta barrel (β barrel). The amino acid composition of the porin β strands are unique in that polar and nonpolar residues alternate along them. This means that the nonpolar residues face outward so as to interact with the nonpolar lipids of outer membrane, whereas the polar residues face inwards into the center of the beta barrel to create the aqueous channel. The specific amino acids in the channel determine the specificity of the porin to different molecules.

The model's confidence: 0.85. Finding RW-7815 / 1041856 / #7, read from revision 1374052512 of Porin (protein).

The model's reasoning: This states the paper's claimed direct conversion of adult skin fibroblasts into neurons as fact; data and image fabrication undermine that finding.

The whole list item: Bypassing stem cells, scientists have made neurons directly from human skin. (Cell)

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-221 / 30485345 / #1, read from revision 1369946876 of 2011 in science.

The model's reasoning: The item presents SCIPIO's reported therapeutic benefits as credible results, although the paper was retracted for unreliable data and conclusions; no other supporting citations were supplied.

The whole list item: A study of heart failure patients treated with their own stem cells has achieved striking results and could result in the biggest breakthrough in a generation. (Daily Telegraph) (The Lancet)

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-20119 / 30485345 / #1, read from revision 1369946876 of 2011 in science.

The model's reasoning: TRIM28 is TIF1beta; the list asserts the physical interaction with C/EBPbeta reported in the retracted paper, while the 1998 paper plausibly independently supports that interaction.

Other citations on it:

• Chang CJ, Chen YL, Lee SC (October 1998). "Coactivator TIF1beta interacts with transcription factor C/EBPbeta and glucocorticoid receptor to induce alpha1-acid glycoprotein gene expression". Molecular and Cellular Biology. 18 (10): 5880–5887. doi:10.1128/mcb.18.10.5880. PMC 109174. PMID 9742105.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-4666 / 14071313 / #1, read from revision 1362595177 of CEBPB.

The model's reasoning: This echoes the withdrawn review's conclusion of poor evidence for dietary efficacy, but its later date may indicate a separate updated source that an editor must check.

Where the model found it: The theory further states that removing opiate precursors from a child's diet may allow time for these behaviors to cease, and neurological development in very young children to resume normally. As of 2021, reliable studies have not demonstrated the benefit of gluten-free diets in the treatment of autism. In the subset of people who have gluten sensitivity there is limited evidence that suggests that a gluten-free diet may improve some autistic behaviors.

The model's confidence: 0.86. Finding RW-30715 / 4071231 / #6, read from revision 1376296858 of Causes of autism.

The model's reasoning: This presents the review's finding that no surgical technique demonstrated superior pain relief, although the review also cautioned that equivalence was not established.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-30945 / 40934678 / #1, read from revision 1378638044 of Osteoarthritis at the base of the thumb.

The model's reasoning: This repeats the cited review's comparison of trapeziectomy variants; no independent source is visible here, so an editor should check the paragraph's sources.

Where the model found it: Some physicians believe that combining LR with TI will help maintain gap between the metacarpal and the scaphoid. And that doing so will improve comfort and capability. Keep in mind that these aspects of the rationale are not supported by experimental evidence. The evidence suggests that all of these procedures have comparable long-term results.

The model's confidence: 0.91. Finding RW-30945 / 40934678 / #2, read from revision 1378638044 of Osteoarthritis at the base of the thumb.

The model's reasoning: This matches the review's finding of insufficient evidence for advantages of joint implants over trapeziectomy; an editor should check whether this paragraph has an independent source.

The model's confidence: 0.85. Finding RW-30945 / 40934678 / #3, read from revision 1378638044 of Osteoarthritis at the base of the thumb.

The model's reasoning: The paper reports 14-day survival after experimental oesophageal transplantation in rats, but the row's named researcher and dates do not match it, so the source of this outcome needs checking.

The whole table row: 21 Esophagus transplant (experimental) Dr. Francisco T. A. Leite et al. Unnamed Deceased donor / synthetic scaffold Experimental in early 2000s; not standard ~2003–2011 Short-term survival NIH

Other citations on it:

• Jensen, Todd; Blanchette, Alex; Vadasz, Stephanie; Dave, Apeksha; Canfarotta, Michael; Sayej, Wael N.; Finck, Christine (July 2015). "Biomimetic and synthetic esophageal tissue engineering". Biomaterials. 57: 133–141. doi:10.1016/j.biomaterials.2015.04.004. ISSN 1878-5905. PMID 25916501.
• Kim, In Gul; Wu, Yanru; Park, Su A.; Choi, Ji Suk; Kwon, Seong Keun; Choi, Seung Hong; Jung, Kyeong Cheon; Shin, Jung-Woog; Chung, Eun-Jae (May 2023). "Assessment of Esophageal Reconstruction via Bioreactor Cultivation of a Synthetic Scaffold in a Canine Model". Clinical and Experimental Otorhinolaryngology. 16 (2): 165–176. doi:10.21053/ceo.2022.01522. ISSN 1976-8710. PMC 10208848. PMID 36652920.
• La Francesca, Saverio; Aho, Johnathon M.; Barron, Matthew R.; Blanco, Ellen W.; Soliman, Sherif; Kalenjian, Lena; Hanson, Ariel D.; Todorova, Elisaveta; Marsh, Matthew; Burnette, Ka Lia; DerSimonian, Harout; Odze, Robert D.; Wigle, Dennis A. (2018-12-01). "Long-term regeneration and remodeling of the pig esophagus after circumferential resection using a retrievable synthetic scaffold carrying autologous cells". Scientific Reports. 8 (1) 4123. Bibcode:2018NatSR...8.4123L. doi:10.1038/s41598-018-22401-x. ISSN 2045-2322. PMC 5841275. PMID 29515136.
• Tan, J. Y.; Chua, C. K.; Leong, K. F.; Chian, K. S.; Leong, W. S.; Tan, L. P. (January 2012). "Esophageal tissue engineering: an in-depth review on scaffold design". Biotechnology and Bioengineering. 109 (1): 1–15. Bibcode:2012BiotB.109....1T. doi:10.1002/bit.23323. ISSN 1097-0290. PMID 21915849.

The reference is tagged {{Retracted}}.

The model's confidence: 0.88. Finding RW-6433 / 8338783 / #1, read from revision 1378522902 of List of organ transplant donors and recipients.

The model's reasoning: This duplicate experimental oesophagus row repeats the cited row's survival claim, potentially drawing on the retracted rat study; an editor must check its own sources, especially for the human outcomes and graft-necrosis claim.

Where the model found it: Esophagus (cadaveric transplant – experimental) Dr. Francisco T. A. Leite et al. Unnamed (human and animal models) Deceased donor / synthetic scaffold Experimental donor/synthetic esophagus transplants began in early 2000s; mixed results, not standard practice ~2003–2011 (experimental phase) Short-term survival in some animal/human trials; long-term failure due to graft necrosis NIH

The model's confidence: 0.86. Finding RW-6433 / 8338783 / #2, read from revision 1378522902 of List of organ transplant donors and recipients.

The model's reasoning: The yield increase is a finding of the retracted paper; the accompanying retraction notice is not independent support for that finding.

The sentence the footnote ends: Although it is not clear exactly how p19 works to suppress RNA silencing, studies have shown that transiently expressed proteins in Nicotiana benthamiana leaves have an up to 50-fold higher yield when coinfiltrated with TBSV p19.

Other citations on it:

• "Retraction: 'An enhanced transient expression system in plants based on suppression of gene silencing by the p19 protein of tomato bushy stunt virus'". The Plant Journal. 84 (4): 846. 2015-11-01. Bibcode:2015PlJ....84..846.. doi:10.1111/tpj.13066. ISSN 1365-313X. PMID 27170951.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-2873 / 7090906 / #1, read from revision 1321932571 of Agroinfiltration.

The model's reasoning: The comparative efficacy claim matches the review's conclusion that granisetron and ondansetron have similar efficacy; the review was withdrawn as out of date.

The sentence the footnote ends: It may be used for chemotherapy-induced nausea and vomiting and appears to work about the same as ondansetron.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-32709 / 429616 / #1, read from revision 1335737669 of Granisetron.

The model's reasoning: The citation supports this specific origin claim, but the title alone does not establish whether it is the paper's own conclusion or historical background.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.90. Finding RW-68812 / 32465321 / #1, read from revision 1362341070 of Tauco.

The model's reasoning: The review supports the characterization of trial evidence as inconclusive, not the sentence’s later claim of established efficacy; its replacement should be checked.

The sentence the footnote ends: Although the results of early trials were inconclusive as the therapy advanced from its initial applications in the mid-1990s, meta-analytic reviews suggested CBT to be an effective treatment for the psychotic symptoms of schizophrenia.

Other citations on it:

• Wykes T, Steel C, Everitt B, Tarrier N (May 2008). "Cognitive behavior therapy for schizophrenia: effect sizes, clinical models, and methodological rigor". Schizophrenia Bulletin. 34 (3): 523–37. doi:10.1093/schbul/sbm114. PMC 2632426. PMID 17962231.
• Zimmermann G, Favrod J, Trieu VH, Pomini V (September 2005). "The effect of cognitive behavioral treatment on the positive symptoms of schizophrenia spectrum disorders: a meta-analysis". Schizophrenia Research. 77 (1): 1–9. doi:10.1016/j.schres.2005.02.018. PMID 16005380. S2CID 31560136.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-33452 / 12688516 / #1, read from revision 1377409896 of Management of schizophrenia.

The model's reasoning: This presents the paper's reported immunological finding as a result; the surrounding misconduct allegations do not explicitly say the finding was refuted or the paper retracted.

Where the model says the article warns about this paper:

• In the same section (Investigation and exoneration): Margot O'Toole, a researcher in Imanishi-Kari's lab, claimed she could not reproduce some of the experiments in the paper and accused Imanishi-Kari of fabricating the data.
• In the same section (Investigation and exoneration): Largely on the basis of these findings, NIH's fraud unit, then called the Office of Scientific Integrity, accused Dr. Imanishi-Kari in 1991 of falsifying data and recommended she be barred from receiving research grants for 10 years.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.88. Finding RW-194 / 3029772 / #1, read from revision 1356628743 of Thereza Imanishi-Kari.

The model's reasoning: The title identifies a review of Chinese herbal medicine for eczema, but without its abstract the claimed uncertainty about efficacy cannot be confirmed as its conclusion; the other systematic review plausibly supports the claim.

Other citations on it:

• Shi ZF, Song TB, Xie J, Yan YQ, Du YP (2017). "The Traditional Chinese Medicine and Relevant Treatment for the Efficacy and Safety of Atopic Dermatitis: A Systematic Review and Meta-Analysis of Randomized Controlled Trials". Evidence-Based Complementary and Alternative Medicine. 2017 6026434. doi:10.1155/2017/6026434. PMC 5497608. PMID 28713436.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.86. Finding RW-32854 / 975668 / #1, read from revision 1350699885 of Antipruritic.

The model's reasoning: The item states the paper's description of the abnormal RCD II lymphocyte phenotype, including its specific marker pattern.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-41496 / 12228277 / #1, read from revision 1343052098 of Gluten-sensitive enteropathy–associated conditions.

The model's reasoning: This closely parallels the paper's definition of RCD and its requirement to exclude alternative diagnoses; an editor must check whether this paragraph has an independent source.

Where the model found it: CD is associated with two grades of disease linked precancerous states. This condition is known as refractory coeliac disease (RCD), defined as malabsorption due to gluten-related enteropathy (villous atrophy or elevated intraepitheal lymphocytes) after initial or subsequent failure of a strict gluten-free diet (usually 1 year) and after exclusion of any disorder mimicking coeliac disease.

The model's confidence: 0.86. Finding RW-41496 / 12228277 / #2, read from revision 1343052098 of Gluten-sensitive enteropathy–associated conditions.

The model's reasoning: These treatment claims repeat the paper's discussion of azathioprine, prednisone, and elemental diets; local footnotes are unavailable, so independent sourcing needs checking.

Where the model found it: RCD 1 involves precancerous tissues in which transformed T-cells continue to produce a response even though gluten is no longer present. Some RCD1 patients have been treated successfully with immunosuppressants (azathioprine, prednisone) when caught early. Elemental diet (proteins digested to amino acids) seems to be an effective alternative diet, indicating other proteins are stimulating the IEL. 5 year viability is high when treated. DQ representation is similar to non-RCD coeliacs.

The model's confidence: 0.90. Finding RW-41496 / 12228277 / #3, read from revision 1343052098 of Gluten-sensitive enteropathy–associated conditions.

The model's reasoning: The lymphoma risk and poorer RCD II prognosis echo the paper's reported outcomes, although the precise survival comparison is not in the supplied abstract; an editor must check the item's own sources.

Where the model found it: Clonal T-cell expansion in RCD2 is not manageable with steroids (see: RCD 1) and sometimes manageable with chemotherapeutic drugs, however, more aggressive therapies seem more affective. A high percentage of RCD 2 patients spontaneously develop lymphoma (EATL), the 5 year survival rate is markedly lower than RCD1 but higher than lymphoma. DQ2.5/DQ2 individuals are more frequently found.

The model's confidence: 0.80. Finding RW-41496 / 12228277 / #4, read from revision 1343052098 of Gluten-sensitive enteropathy–associated conditions.

The model's reasoning: The caption presents the paper's phylogenetic analysis without a local retraction warning; this ancillary result needs independent sourcing.

Where the model says the article warns about this paper:

• In the opening paragraphs: Subsequent independent studies published in 2012 found no detectable arsenate in the DNA of GFAJ-1, refuted the claim, and demonstrated that GFAJ-1 is simply an arsenate-resistant, phosphate-dependent organism.
• In the opening paragraphs: The paper was retracted by the Science journal in 2025.
• Later (Biochemistry): After finding and addressing further issues (ionic strength, pH and the use of glass tubes instead of polypropylene) Redfield found that arsenate marginally stimulated growth, but didn't affect the final densities of the cultures, unlike what was claimed.
• Later (Criticism of Science study): In January 2012 a group of researchers led by Rosie Redfield at the University of British Columbia analyzed the DNA of GFAJ-1 using liquid chromatography–mass spectrometry and could not detect any arsenic, which Redfield calls a "clear refutation" of the original paper's findings.
• Later (Criticism of Science study): The paper was retracted in 2025 with "no evidence of misconduct by researchers".

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.92. Finding RW-66335 / 29856330 / #3, read from revision 1374726997 of GFAJ-1.

The model's reasoning: The citation supports the paper's phylogenetic result, which is presented as valid and may be independently sourceable.

The sentence the footnote ends: Although the authors produced a cladogram in which the strain is nested among members of Halomonas, including H. alkaliphila and H. venusta, they did not explicitly assign the strain to that genus.

Where the model says the article warns about this paper:

• In the opening paragraphs: Subsequent independent studies published in 2012 found no detectable arsenate in the DNA of GFAJ-1, refuted the claim, and demonstrated that GFAJ-1 is simply an arsenate-resistant, phosphate-dependent organism.
• In the opening paragraphs: The paper was retracted by the Science journal in 2025.
• Later (Biochemistry): After finding and addressing further issues (ionic strength, pH and the use of glass tubes instead of polypropylene) Redfield found that arsenate marginally stimulated growth, but didn't affect the final densities of the cultures, unlike what was claimed.
• Later (Criticism of Science study): In January 2012 a group of researchers led by Rosie Redfield at the University of British Columbia analyzed the DNA of GFAJ-1 using liquid chromatography–mass spectrometry and could not detect any arsenic, which Redfield calls a "clear refutation" of the original paper's findings.
• Later (Criticism of Science study): The paper was retracted in 2025 with "no evidence of misconduct by researchers".

Other citations on it:

• Bortman, Henry (2 December 2010). "Thriving on arsenic". Astrobiology Magazine (NASA). Retrieved 11 December 2010.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.95. Finding RW-66335 / 29856330 / #4, read from revision 1374726997 of GFAJ-1.

The model's reasoning: The sentence combines a statement about the authors' naming choice with an assessment of species criteria whose support requires source inspection.

Where the model says the article warns about this paper:

• In the opening paragraphs: Subsequent independent studies published in 2012 found no detectable arsenate in the DNA of GFAJ-1, refuted the claim, and demonstrated that GFAJ-1 is simply an arsenate-resistant, phosphate-dependent organism.
• In the opening paragraphs: The paper was retracted by the Science journal in 2025.
• Later (Biochemistry): After finding and addressing further issues (ionic strength, pH and the use of glass tubes instead of polypropylene) Redfield found that arsenate marginally stimulated growth, but didn't affect the final densities of the cultures, unlike what was claimed.
• Later (Criticism of Science study): In January 2012 a group of researchers led by Rosie Redfield at the University of British Columbia analyzed the DNA of GFAJ-1 using liquid chromatography–mass spectrometry and could not detect any arsenic, which Redfield calls a "clear refutation" of the original paper's findings.
• Later (Criticism of Science study): The paper was retracted in 2025 with "no evidence of misconduct by researchers".

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.76. Finding RW-66335 / 29856330 / #7, read from revision 1374726997 of GFAJ-1.

The model's reasoning: This states the original growth result as fact without a local retraction or refutation warning.

Where the model says the article warns about this paper:

• In the opening paragraphs: Subsequent independent studies published in 2012 found no detectable arsenate in the DNA of GFAJ-1, refuted the claim, and demonstrated that GFAJ-1 is simply an arsenate-resistant, phosphate-dependent organism.
• In the opening paragraphs: The paper was retracted by the Science journal in 2025.
• In the same section (Biochemistry): After finding and addressing further issues (ionic strength, pH and the use of glass tubes instead of polypropylene) Redfield found that arsenate marginally stimulated growth, but didn't affect the final densities of the cultures, unlike what was claimed.
• Later (Criticism of Science study): In January 2012 a group of researchers led by Rosie Redfield at the University of British Columbia analyzed the DNA of GFAJ-1 using liquid chromatography–mass spectrometry and could not detect any arsenic, which Redfield calls a "clear refutation" of the original paper's findings.
• Later (Criticism of Science study): The paper was retracted in 2025 with "no evidence of misconduct by researchers".

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-66335 / 29856330 / #8, read from revision 1374726997 of GFAJ-1.

The model's reasoning: These quantitative measurements are presented as established results and supported only by the retracted experiment.

The sentence the footnote ends: This phosphorus content was also only about one tenth of the cells' average arsenic content (0.19 ± 0.25% by dry weight).

Where the model says the article warns about this paper:

• In the opening paragraphs: Subsequent independent studies published in 2012 found no detectable arsenate in the DNA of GFAJ-1, refuted the claim, and demonstrated that GFAJ-1 is simply an arsenate-resistant, phosphate-dependent organism.
• In the opening paragraphs: The paper was retracted by the Science journal in 2025.
• In the same section (Biochemistry): After finding and addressing further issues (ionic strength, pH and the use of glass tubes instead of polypropylene) Redfield found that arsenate marginally stimulated growth, but didn't affect the final densities of the cultures, unlike what was claimed.
• Later (Criticism of Science study): In January 2012 a group of researchers led by Rosie Redfield at the University of British Columbia analyzed the DNA of GFAJ-1 using liquid chromatography–mass spectrometry and could not detect any arsenic, which Redfield calls a "clear refutation" of the original paper's findings.
• Later (Criticism of Science study): The paper was retracted in 2025 with "no evidence of misconduct by researchers".

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-66335 / 29856330 / #9, read from revision 1374726997 of GFAJ-1.

The model's reasoning: The sentence uses measurements from the retracted study as evidence, despite their contrast with its headline claim.

Where the model says the article warns about this paper:

• In the opening paragraphs: Subsequent independent studies published in 2012 found no detectable arsenate in the DNA of GFAJ-1, refuted the claim, and demonstrated that GFAJ-1 is simply an arsenate-resistant, phosphate-dependent organism.
• In the opening paragraphs: The paper was retracted by the Science journal in 2025.
• In the same section (Biochemistry): After finding and addressing further issues (ionic strength, pH and the use of glass tubes instead of polypropylene) Redfield found that arsenate marginally stimulated growth, but didn't affect the final densities of the cultures, unlike what was claimed.
• Later (Criticism of Science study): In January 2012 a group of researchers led by Rosie Redfield at the University of British Columbia analyzed the DNA of GFAJ-1 using liquid chromatography–mass spectrometry and could not detect any arsenic, which Redfield calls a "clear refutation" of the original paper's findings.
• Later (Criticism of Science study): The paper was retracted in 2025 with "no evidence of misconduct by researchers".

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-66335 / 29856330 / #10, read from revision 1374726997 of GFAJ-1.

The model's reasoning: The growth-rate and cell-volume measurements are asserted as facts from the retracted study without independent support.

The sentence the footnote ends: The phosphate-starved bacteria had an intracellular volume 1.5 times normal; the greater volume appeared to be associated with the appearance of large "vacuole-like regions".

Where the model says the article warns about this paper:

• In the opening paragraphs: Subsequent independent studies published in 2012 found no detectable arsenate in the DNA of GFAJ-1, refuted the claim, and demonstrated that GFAJ-1 is simply an arsenate-resistant, phosphate-dependent organism.
• In the opening paragraphs: The paper was retracted by the Science journal in 2025.
• In the same section (Biochemistry): After finding and addressing further issues (ionic strength, pH and the use of glass tubes instead of polypropylene) Redfield found that arsenate marginally stimulated growth, but didn't affect the final densities of the cultures, unlike what was claimed.
• Later (Criticism of Science study): In January 2012 a group of researchers led by Rosie Redfield at the University of British Columbia analyzed the DNA of GFAJ-1 using liquid chromatography–mass spectrometry and could not detect any arsenic, which Redfield calls a "clear refutation" of the original paper's findings.
• Later (Criticism of Science study): The paper was retracted in 2025 with "no evidence of misconduct by researchers".

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.98. Finding RW-66335 / 29856330 / #11, read from revision 1374726997 of GFAJ-1.

The model's reasoning: The proposed stabilization mechanism is reported without explicit refutation or retraction; the other sources may support background or attribution, not validate the mechanism.

Where the model says the article warns about this paper:

• In the opening paragraphs: Subsequent independent studies published in 2012 found no detectable arsenate in the DNA of GFAJ-1, refuted the claim, and demonstrated that GFAJ-1 is simply an arsenate-resistant, phosphate-dependent organism.
• In the opening paragraphs: The paper was retracted by the Science journal in 2025.
• In the same section (Biochemistry): After finding and addressing further issues (ionic strength, pH and the use of glass tubes instead of polypropylene) Redfield found that arsenate marginally stimulated growth, but didn't affect the final densities of the cultures, unlike what was claimed.
• Later (Criticism of Science study): In January 2012 a group of researchers led by Rosie Redfield at the University of British Columbia analyzed the DNA of GFAJ-1 using liquid chromatography–mass spectrometry and could not detect any arsenic, which Redfield calls a "clear refutation" of the original paper's findings.
• Later (Criticism of Science study): The paper was retracted in 2025 with "no evidence of misconduct by researchers".

Other citations on it:

• Mulyana, Ilham; Satoko, Nakanomori; Takahiro, Kihara; Ayaka, Hokamura; Matsusaki, Hiromi; Takeharu, Tsuge; Kouhei, Mizuno (2014). "Characterization of Polyhydroxyalkanoate Synthases from Halomonas sp. O-1 and Halomonas elongata DSM2581: Site-Directed Mutagenesis and Recombinant Expression". Polymer Degradation and Stability. 109: 416–429. doi:10.1016/j.polymdegradstab.2014.04.024.
• Bortman, Henry (2 December 2010). "Thriving on arsenic". Astrobiology Magazine (NASA). Retrieved 11 December 2010.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.91. Finding RW-66335 / 29856330 / #13, read from revision 1374726997 of GFAJ-1.

The model's reasoning: The linked title repeats the publicized arsenic-based-life finding associated with this paper; an editor should check the linked source and whether this historical link is intentionally retained.

Where the model found it: NASA.gov: "NASA-Funded Research Discovers Life Built With Toxic Chemical" Archived 28 August 2011 at the Wayback Machine — (2 December 2010)

Where the model says the article warns about this paper:

• In the opening paragraphs: Subsequent independent studies published in 2012 found no detectable arsenate in the DNA of GFAJ-1, refuted the claim, and demonstrated that GFAJ-1 is simply an arsenate-resistant, phosphate-dependent organism.
• In the opening paragraphs: The paper was retracted by the Science journal in 2025.
• Earlier (Biochemistry): After finding and addressing further issues (ionic strength, pH and the use of glass tubes instead of polypropylene) Redfield found that arsenate marginally stimulated growth, but didn't affect the final densities of the cultures, unlike what was claimed.
• Earlier (Criticism of Science study): In January 2012 a group of researchers led by Rosie Redfield at the University of British Columbia analyzed the DNA of GFAJ-1 using liquid chromatography–mass spectrometry and could not detect any arsenic, which Redfield calls a "clear refutation" of the original paper's findings.
• Earlier (Criticism of Science study): The paper was retracted in 2025 with "no evidence of misconduct by researchers".

The model's confidence: 0.78. Finding RW-66335 / 29856330 / #14, read from revision 1374726997 of GFAJ-1.

The model's reasoning: These sentences report the paper’s persistence and household-transmission findings; “purported” does not explicitly tell readers here that the findings were fabricated or the paper retracted.

The sentence the footnote ends: LaCour and Green (2014) also found strong evidence of within-household transmission of opinion change, but only in the wake of conversations with gay canvassers.

Where the model says the article warns about this paper:

• In the opening paragraphs: "When Contact Changes Minds: An Experiment on Transmission of Support for Gay Equality" is a fraudulent article by then-UCLA political science graduate student Michael LaCour and Columbia University political science professor Donald Green. The article was published in the academic journal Science in December 2014, and retracted in May 2015 after it emerged that the data in the study had been forged by LaCour.
• Later (Falsified data): The "When contact changes minds" study was discredited after a critique by David Broockman, Joshua Kalla, and Peter Aronow on May 19, 2015, titled "Irregularities in LaCour (2014)", concluded that the data had been falsified and no data had been collected.
• Later (Falsified data): The "Irregularities" paper also identified the likely method by which LaCour had forged the data.
• Later (Falsified data): Additional evidence of the study's fictitiousness later emerged, such as evidence that LaCour had tried to retroactively claim that the study had been pre-registered using falsified documents.
• Later (Falsified data): LaCour's coauthor, Donald Green, subsequently requested that the paper be retracted because of “irregularities” in the paper's data and LaCour's failure to give him the raw data on which the paper was based.
• Later (Falsified data): On May 28, 2015, the study was retracted by Science, on the basis that incentives to participate in the survey on which the study was based had been misrepresented, that sponsors had been falsely identified, that the authors could not produce the original data, and (citing the findings of Broockman et al.) that there were "statistical irregularities."
• Later (Falsified data): After the retraction, the Carnegie Corporation of New York rescinded Donald Green's 2015 Andrew Carnegie Fellowship, revoking a $200,000 award to support Donald Green's research, and Princeton University rescinded an assistant professorship that had been offered to LaCour.
• Later (Falsified data): The journal Science retracted the article with the concurrence of Green while LaCour did not agree to the retraction.
• Later (Popular press): Conservative pro-gay-marriage columnist S.E. Cupp wrote that, "The doctored study will only encourage the perception that advocates are going too far."
• Later (Follow-up study): Following their debunking of When Contact Changes Minds, Broockman and Kalla (2016) conducted an experiment like the one purported to have been conducted by LaCour and Green.
• Later (Follow-up study): However, contrary to one of LaCour's claims, the identity of the canvasser did not seem to matter; both transgender and cisgender canvassers had a similar effect.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-520 / 46759922 / #1, read from revision 1369324476 of When Contact Changes Minds.

The model's reasoning: This states the review's conclusions about early seizure prevention and the lack of demonstrated prevention of late seizures; the other reviews plausibly support both claims.

Other citations on it:

• Pagni CA, Zenga F (2005). "Posttraumatic epilepsy with special emphasis on prophylaxis and prevention". Acta Neurochirurgica. Acta Neurochirurgica Supplementum. 93: 27–34. doi:10.1007/3-211-27577-0_3. ISBN 978-3-211-24150-9. PMID 15986723.{{cite journal}}: CS1 maint: periodical has ISBN (link)
• Garga N, Lowenstein DH (2006). "Posttraumatic Epilepsy: A Major Problem in Desperate Need of Major Advances". Epilepsy Currents. 6 (1): 1–5. doi:10.1111/j.1535-7511.2005.00083.x. PMC 1363374. PMID 16477313.
• D'Ambrosio R, Perucca E (2004). "Epilepsy after head injury". Current Opinion in Neurology. 17 (6): 731–735. doi:10.1097/00019052-200412000-00014. PMC 2672045. PMID 15542983.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-33075 / 15731985 / #1, read from revision 1354071677 of Post-traumatic epilepsy.

The model's reasoning: This presents the review's early-seizure prevention conclusion as established evidence, with no other citation on the sentence; retract-and-replace calls for checking the replacement review.

The sentence the footnote ends: Studies did show that antiepileptics prevented seizures occurring within a week after injury, and in 1995 the task force of the Brain Trauma Foundation published a recommendation suggesting their use to protect against seizures early after trauma.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-33075 / 15731985 / #3, read from revision 1354071677 of Post-traumatic epilepsy.

The model's reasoning: The prevention claim summarizes the review's conclusion cited in the Prevention section, although this lead paragraph may have independent sourcing that an editor should check.

Where the model found it: Diagnostic measures include electroencephalography (EEG) and brain imaging techniques such as magnetic resonance imaging, but these are not totally reliable. Antiepileptic drugs do not prevent the development of PTE after head injury, but may be used to treat the condition if it does occur. When medication does not work to control the seizures, surgery may be needed. Modern surgical techniques for PTE have their roots in the 19th century, but trepanation (cutting the skull to make a hole) may have been used for the condition in ancient cultures.

The model's confidence: 0.94. Finding RW-33075 / 15731985 / #4, read from revision 1354071677 of Post-traumatic epilepsy.

The model's reasoning: This conclusion draws on the review's finding of poor evidence for efficacy; the other review plausibly supports it independently.

The sentence the footnote ends: While some proponents claim such diets can alleviate or cure autism symptoms—often relying on anecdotal or non-scientific evidence—there is no conclusive scientific support for these claims.

Other citations on it:

• Christison GW, Ivany K; Ivany (2006). "Elimination diets in autism spectrum disorders: any wheat amidst the chaff?". J Dev Behav Pediatr. 27 (2 Suppl 2): S162–71. doi:10.1097/00004703-200604002-00015. PMID 16685183.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-30715 / 195522 / #1, read from revision 1377992970 of Gluten-free, casein-free diet.

The model's reasoning: The sentence reports the review's conclusion without noting its withdrawal; withdrawal as out of date calls for an updated source rather than removal on grounds of invalid findings.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-30715 / 195522 / #3, read from revision 1377992970 of Gluten-free, casein-free diet.

The model's reasoning: The item presents the review's evidence assessment as an official position without noting withdrawal; the additional review plausibly supports the assessment.

The whole list item: Cochrane Library – Gluten and Casein-free diets in autism spectrum disorder (2008). The Cochrane review found that while relatively commonly used the evidence to support the diets use in children with autism was poor. All studies as of 2006 had issues with them.

Other citations on it:

• Christison GW, Ivany K; Ivany (2006). "Elimination diets in autism spectrum disorders: any wheat amidst the chaff?". J Dev Behav Pediatr. 27 (2 Suppl 2): S162–71. doi:10.1097/00004703-200604002-00015. PMID 16685183.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-30715 / 195522 / #4, read from revision 1377992970 of Gluten-free, casein-free diet.

The model's reasoning: This repeats the retracted paper's finding that player-guided optimization outperforms numerical optimization; an editor should check whether this paragraph has an independent supporting source.

Where the model found it: In Quantum Moves, the atomic wave function is represented as a sloshy liquid in an energy potential well created by the optical tweezers. Players control the depth and the horizontal location of the well, simulating the path on the optical tweezers. The wave function reacts to changes in the potential function as dictated by the Schrödinger equation leading to sloshing seen by the players. Players are asked to move the well without sloshing the atomic wave function too much. A path created by a player maps one-to-one to a solution of the Schrödinger equation. Top results of the game play are then used to provide guidance into the algorithm's search space, resulting in solutions superior to those found by the algorithm alone.

Where the model says the article warns about this paper:

• Later (History): The article was later retracted (see below).
• Later (History): He subsequently discovered that the authors of the original Nature paper had made a sign error in their implementation of the benchmarked optimization algorithm, which led to the retraction of the Nature paper in July 2020.
• Later (Controversy and retraction): Controversy and retraction
• Later (Controversy and retraction): In 2020 the Nature article where the findings were presented was retracted due to major errors in calculations, deeming the results of the article false and the game untrustworthy.
• Later (Controversy and retraction): The coordinator of the project continued presenting the data as true until 2020 where an internal investigation from the University of Aarhus discovered there were problems in the way the equations were implemented, resulting in a mistake which deemed the findings false as other scientists have claimed since the release of the article in 2016.

The model's confidence: 0.96. Finding RW-23835 / 41008786 / #3, read from revision 1357721035 of Quantum Moves.

The model's reasoning: The paragraph reports the Japanese paper's claimed diazaannulene synthesis without a local retraction or refutation warning; the other citation supports the American claim, not the Japanese result.

The sentence the footnote ends: In 2006 and again in 2007 the Zincke reaction was rediscovered by a research group from Japan and a group from the USA.

Where the model says the article warns about this paper:

• In the same section (2007 rediscovery): In a letter to Angewandte Chemie, the German chemist Manfred Christl pointed out not only that the alleged new chemistry was in fact 100-year-old Zincke chemistry but also that the proposed structure for the reaction product was not the 12 membered ring but the 6 membered pyridinium salt (structure 2).
• In the same section (2007 rediscovery): As of December 2007 the Japanese group retracted its paper in Organic Letters due to uncertainties regarding what products are formed in the reaction described

Other citations on it:

• Shi, Lei; Lundberg, Dan; Musaev, Djamaladdin G.; Menger, Fredric M. (2007). "[12]Annulene Gemini Surfactants: Structure and Self-Assembly". Angewandte Chemie International Edition. 46 (31): 5889–5891. doi:10.1002/anie.200702140. PMID 17615610.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.97. Finding RW-1811 / 7974411 / #1, read from revision 1328566283 of Zincke reaction.

The model's reasoning: This directly reproduces the retracted paper's mortality and hormone-dependent findings without a local retraction or refutation warning; an editor should check any local footnotes.

Where the model says the article warns about this paper:

• In the opening paragraphs: The Séralini affair was the controversy surrounding the publication, retraction, and republication of a journal article by French molecular biologist Gilles-Éric Séralini.
• In the opening paragraphs: Scientists and regulatory agencies subsequently concluded that the study's design was flawed and its findings unsubstantiated.
• In the opening paragraphs: Following widespread criticism by scientists, Food and Chemical Toxicology retracted the paper in November 2013 after the authors refused to withdraw it. The editor-in-chief said that the article was retracted because its data were inconclusive and its conclusions unreliable.
• Later (Scientific evaluation): Many said that Séralini's conclusions were impossible to justify due to experimental design issues or lack of statistical power of the study.
• Later (Scientific evaluation): The study shows both shortcomings in study design and in the presentation of the collected data. This means that the conclusions drawn by the authors are not supported by the available data.
• Later (Scientific evaluation): The study as reported by Séralini et al. was found to be inadequately designed, analysed and reported...The study as described by Séralini et al. does not allow giving weight to their results and conclusions as published. Conclusions cannot be drawn on the difference in tumour incidence between treatment groups on the basis of the design, the analysis and the results as reported.
• Later (Scientific evaluation): However, given this study presents serious deficiencies in the protocol, the procedures and the interpretation of the results, the SFPT cannot support any of the scientific claims drawn by the authors, and any relevance for human risk assessment.
• Later (Scientific evaluation): Using Seralini's published numerical data, the review found no significant effects on animal health after analysis with statistical tests.
• Later (Scientific evaluation): The finding that "in females, all the treated groups died 2–3 times more than controls" was not statistically significant. The highest mortality was observed for the group of female rats fed 22% genetically modified maize. This difference was not statistically significant. Seralini also originally claimed males in groups fed 22% and 33% genetically modified maize had three times lower mortality than controls, but this was also not statistically significant. The findings of liver necrosis and mammary tumors were also not significant.
• Later (Scientific evaluation): A 2017 study found that since it was retracted, Seralini et al. (2012) had been cited 60 times after it was retracted, and that more of these citations were negative (39%) than were positive (26%).
• Later (Responses to criticism): Other Séralini supporters criticized the retraction of the study, concluding the response was a product of industry-driven campaign and regard this as a concerning example of industry interference in the scientific process.
• Later (Media): [Seralini] has crossed the line from merely performing and reporting flawed experiments to committing gross scientific misconduct and attempting fraud.
• Later (Retraction): Retraction
• Later (Retraction): In November 2013, Elsevier announced that FCT was retracting the paper, after the authors refused to withdraw it.
• Later (Retraction): After an in-depth look at the study's raw data, no definitive conclusions could be reached regarding the role of either NK603 or glyphosate in overall mortality or tumor rates, given the high incidence of tumors in Sprague-Dawley rats and the small sample size. Normal variance could not be excluded as the cause of the results.
• Later (Retraction): The data are inconclusive, therefore the claim (i.e., conclusion) that Roundup Ready maize NK603 and/or the Roundup herbicide have a link to cancer is unreliable.
• Later (Retraction): However, to be very clear, it is the entire paper, with the claim that there is a definitive link between GMO and cancer that is being retracted.
• Later (Retraction): In our analysis, his conclusions cannot be claimed from the data presented in this article.
• Later (Retraction): Séralini and his supporters strongly objected to the retraction, and Séralini himself threatened to sue FCT.
• Later (Retraction): Hayes said in an interview that he did not have a contract with Monsanto when he retracted Séralini's paper, and that his decision to retract it was not influenced by Monsanto at all.
• Later (Republication): was inadequate for evaluation because the number of animals per group was small, the histopathological description of tumours was poor, and incidences of tumours for individual animals were not provided.

The model's confidence: 0.99. Finding RW-3604 / 39331424 / #3, read from revision 1369163903 of Séralini affair.

The model's reasoning: Although attributed to NPR and used critically, this presents a tumor comparison derived from the retracted study as an observation; an editor should check whether the local source independently validates it.

Where the model found it: The Washington Post quoted Marion Nestle, the Paulette Goddard professor in the Department of Nutrition, Food Studies and Public Health at New York University and food safety advocate: "'[I] can't figure it out yet....It's weirdly complicated and unclear on key issues: what the controls were fed, relative rates of tumors, why no dose relationship, what the mechanism might be. I can't think of a biological reason why GMO corn should do this.....So even though I strongly support labeling, I'm skeptical of this study.'" Likewise, Dan Charles, writing for NPR, noted that in the study, rats that ate 33% GM food developed fewer tumors than did those who ate 11% GM food, suggesting the absence of a dose response. University of Calgary Professor Maurice Moloney publicly wondered why the paper contained so many pictures of treated rats with horrific tumors, but no pictures of control group rats.

Where the model says the article warns about this paper:

• In the opening paragraphs: The Séralini affair was the controversy surrounding the publication, retraction, and republication of a journal article by French molecular biologist Gilles-Éric Séralini.
• In the opening paragraphs: Scientists and regulatory agencies subsequently concluded that the study's design was flawed and its findings unsubstantiated.
• In the opening paragraphs: Following widespread criticism by scientists, Food and Chemical Toxicology retracted the paper in November 2013 after the authors refused to withdraw it. The editor-in-chief said that the article was retracted because its data were inconclusive and its conclusions unreliable.
• In the same section (Scientific evaluation): Many said that Séralini's conclusions were impossible to justify due to experimental design issues or lack of statistical power of the study.
• In the same section (Scientific evaluation): The study shows both shortcomings in study design and in the presentation of the collected data. This means that the conclusions drawn by the authors are not supported by the available data.
• In the same section (Scientific evaluation): The study as reported by Séralini et al. was found to be inadequately designed, analysed and reported...The study as described by Séralini et al. does not allow giving weight to their results and conclusions as published. Conclusions cannot be drawn on the difference in tumour incidence between treatment groups on the basis of the design, the analysis and the results as reported.
• In the same section (Scientific evaluation): However, given this study presents serious deficiencies in the protocol, the procedures and the interpretation of the results, the SFPT cannot support any of the scientific claims drawn by the authors, and any relevance for human risk assessment.
• In the same section (Scientific evaluation): Using Seralini's published numerical data, the review found no significant effects on animal health after analysis with statistical tests.
• In the same section (Scientific evaluation): The finding that "in females, all the treated groups died 2–3 times more than controls" was not statistically significant. The highest mortality was observed for the group of female rats fed 22% genetically modified maize. This difference was not statistically significant. Seralini also originally claimed males in groups fed 22% and 33% genetically modified maize had three times lower mortality than controls, but this was also not statistically significant. The findings of liver necrosis and mammary tumors were also not significant.
• In the same section (Scientific evaluation): A 2017 study found that since it was retracted, Seralini et al. (2012) had been cited 60 times after it was retracted, and that more of these citations were negative (39%) than were positive (26%).
• In the same section (Responses to criticism): Other Séralini supporters criticized the retraction of the study, concluding the response was a product of industry-driven campaign and regard this as a concerning example of industry interference in the scientific process.
• In the same section (Media): [Seralini] has crossed the line from merely performing and reporting flawed experiments to committing gross scientific misconduct and attempting fraud.
• Later (Retraction): Retraction
• Later (Retraction): In November 2013, Elsevier announced that FCT was retracting the paper, after the authors refused to withdraw it.
• Later (Retraction): After an in-depth look at the study's raw data, no definitive conclusions could be reached regarding the role of either NK603 or glyphosate in overall mortality or tumor rates, given the high incidence of tumors in Sprague-Dawley rats and the small sample size. Normal variance could not be excluded as the cause of the results.
• Later (Retraction): The data are inconclusive, therefore the claim (i.e., conclusion) that Roundup Ready maize NK603 and/or the Roundup herbicide have a link to cancer is unreliable.
• Later (Retraction): However, to be very clear, it is the entire paper, with the claim that there is a definitive link between GMO and cancer that is being retracted.
• Later (Retraction): In our analysis, his conclusions cannot be claimed from the data presented in this article.
• Later (Retraction): Séralini and his supporters strongly objected to the retraction, and Séralini himself threatened to sue FCT.
• Later (Retraction): Hayes said in an interview that he did not have a contract with Monsanto when he retracted Séralini's paper, and that his decision to retract it was not influenced by Monsanto at all.
• Later (Republication): was inadequate for evaluation because the number of animals per group was small, the histopathological description of tumours was poor, and incidences of tumours for individual animals were not provided.

The model's confidence: 0.90. Finding RW-3604 / 39331424 / #4, read from revision 1369163903 of Séralini affair.

The model's reasoning: This reports an endorsement of the paper's maize–cancer claim without a local refutation or retraction warning; an editor should check the blog citation and its historical context.

Where the model found it: The Guardian's Environmental Blog stated that the study linking GM maize to cancer "must be taken seriously by regulators" and that although it "attracted a torrent of abuse", "it cannot be swept under the carpet". They also noted CRIIGEN's funding of the research and reported Séralini's response: namely, that studies in support of GM food are usually funded by "corporates or by pro-biotech institutions". Proponents of California's GM labeling referendum, Proposition 37, hailed the study.

Where the model says the article warns about this paper:

• In the opening paragraphs: The Séralini affair was the controversy surrounding the publication, retraction, and republication of a journal article by French molecular biologist Gilles-Éric Séralini.
• In the opening paragraphs: Scientists and regulatory agencies subsequently concluded that the study's design was flawed and its findings unsubstantiated.
• In the opening paragraphs: Following widespread criticism by scientists, Food and Chemical Toxicology retracted the paper in November 2013 after the authors refused to withdraw it. The editor-in-chief said that the article was retracted because its data were inconclusive and its conclusions unreliable.
• In the same section (Scientific evaluation): Many said that Séralini's conclusions were impossible to justify due to experimental design issues or lack of statistical power of the study.
• In the same section (Scientific evaluation): The study shows both shortcomings in study design and in the presentation of the collected data. This means that the conclusions drawn by the authors are not supported by the available data.
• In the same section (Scientific evaluation): The study as reported by Séralini et al. was found to be inadequately designed, analysed and reported...The study as described by Séralini et al. does not allow giving weight to their results and conclusions as published. Conclusions cannot be drawn on the difference in tumour incidence between treatment groups on the basis of the design, the analysis and the results as reported.
• In the same section (Scientific evaluation): However, given this study presents serious deficiencies in the protocol, the procedures and the interpretation of the results, the SFPT cannot support any of the scientific claims drawn by the authors, and any relevance for human risk assessment.
• In the same section (Scientific evaluation): Using Seralini's published numerical data, the review found no significant effects on animal health after analysis with statistical tests.
• In the same section (Scientific evaluation): The finding that "in females, all the treated groups died 2–3 times more than controls" was not statistically significant. The highest mortality was observed for the group of female rats fed 22% genetically modified maize. This difference was not statistically significant. Seralini also originally claimed males in groups fed 22% and 33% genetically modified maize had three times lower mortality than controls, but this was also not statistically significant. The findings of liver necrosis and mammary tumors were also not significant.
• In the same section (Scientific evaluation): A 2017 study found that since it was retracted, Seralini et al. (2012) had been cited 60 times after it was retracted, and that more of these citations were negative (39%) than were positive (26%).
• In the same section (Responses to criticism): Other Séralini supporters criticized the retraction of the study, concluding the response was a product of industry-driven campaign and regard this as a concerning example of industry interference in the scientific process.
• In the same section (Media): [Seralini] has crossed the line from merely performing and reporting flawed experiments to committing gross scientific misconduct and attempting fraud.
• Later (Retraction): Retraction
• Later (Retraction): In November 2013, Elsevier announced that FCT was retracting the paper, after the authors refused to withdraw it.
• Later (Retraction): After an in-depth look at the study's raw data, no definitive conclusions could be reached regarding the role of either NK603 or glyphosate in overall mortality or tumor rates, given the high incidence of tumors in Sprague-Dawley rats and the small sample size. Normal variance could not be excluded as the cause of the results.
• Later (Retraction): The data are inconclusive, therefore the claim (i.e., conclusion) that Roundup Ready maize NK603 and/or the Roundup herbicide have a link to cancer is unreliable.
• Later (Retraction): However, to be very clear, it is the entire paper, with the claim that there is a definitive link between GMO and cancer that is being retracted.
• Later (Retraction): In our analysis, his conclusions cannot be claimed from the data presented in this article.
• Later (Retraction): Séralini and his supporters strongly objected to the retraction, and Séralini himself threatened to sue FCT.
• Later (Retraction): Hayes said in an interview that he did not have a contract with Monsanto when he retracted Séralini's paper, and that his decision to retract it was not influenced by Monsanto at all.
• Later (Republication): was inadequate for evaluation because the number of animals per group was small, the histopathological description of tumours was poor, and incidences of tumours for individual animals were not provided.

The model's confidence: 0.92. Finding RW-3604 / 39331424 / #6, read from revision 1369163903 of Séralini affair.

The model's reasoning: This states the paper's claimed arsenic-supported growth as a discovery, without noting its retraction or the failure to reproduce its results.

The whole list item: 2 December – NASA-supported researchers discover the first microorganism known to be able to thrive and reproduce using the toxic chemical arsenic. (NASA) Archived 2017-05-16 at the Wayback Machine (Science)

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-66335 / 25586145 / #1, read from revision 1370259849 of 2010 in science.

The model's reasoning: The paragraph directly quotes the review’s conclusions about GTR’s benefits; its cautions do not disclose that the review was withdrawn as out of date.

The sentence the footnote ends: Therefore, patients and health professionals need to consider the predictability of the technique compared with other methods of treatment before making final decisions on use."

The reference isn't tagged {{Retracted}}.

The model's confidence: 1.00. Finding RW-30611 / 25299326 / #1, read from revision 1320086452 of Chronic periodontitis.

The model's reasoning: The universal claim about task participation falls within the paper's topic, but the title alone cannot establish whether it is a study finding or background.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.85. Finding RW-47382 / 59753398 / #3, read from revision 1374093906 of Stegodyphus dumicola.

The model's reasoning: The title does not establish whether this prey-response mechanism was a finding of the study or background information.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.80. Finding RW-47382 / 59753398 / #4, read from revision 1374093906 of Stegodyphus dumicola.

The model's reasoning: This appears to summarize the task-participation and foraging findings immediately preceding it, but an editor must check whether the comparison has its own source.

The model's confidence: 0.76. Finding RW-47382 / 59753398 / #10, read from revision 1374093906 of Stegodyphus dumicola.

The model's reasoning: The claimed absence of nuclear transfer and introduced transcription factors is a central conclusion of the retracted paper, without a local retraction warning.

Where the model says the article warns about this paper:

• In the opening paragraphs: Over the following months, all scientists who tried to duplicate her results failed, and suspicion arose that Obokata's results were due to error or fraud.
• In the opening paragraphs: On April 1, 2014, RIKEN concluded that Obokata had falsified data to obtain her results. On June 4, 2014, Obokata agreed to retract the papers.
• Later (Investigation into disputed claims): In the months after the two Nature papers were released, all scientists who tried to duplicate Obokata's results failed and suspicion arose that her results were due to error or fraud.
• Later (Investigation into disputed claims): On April 1, RIKEN concluded that Obokata had engaged in "research misconduct", falsifying data on two occasions.
• Later (Investigation into disputed claims): On June 4, 2014, Obokata agreed to retract both the article and the "letter". The article was officially retracted on July 2, 2014.
• Later (Investigation into disputed claims): On September 24, 2015, the RIKEN scientists reported that Obokata's STAP cells came from embryonic stem cell contamination, while on the same day, research groups who had attempted to reproduce the STAP protocol jointly reported that they had found it irreproducible.
• Later (Implications): If the findings had proven to be valid, stimulus-triggered pluripotency cells could have been generated more easily and efficiently than by existing iPS techniques.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-2081 / 41787889 / #1, read from revision 1375988571 of Stimulus-triggered acquisition of pluripotency.

The model's reasoning: This reports the paper's claimed stress-induced reprogramming; NHS coverage plausibly supports the reported protocol but is not independent validation.

Where the model says the article warns about this paper:

• In the opening paragraphs: Over the following months, all scientists who tried to duplicate her results failed, and suspicion arose that Obokata's results were due to error or fraud.
• In the opening paragraphs: On April 1, 2014, RIKEN concluded that Obokata had falsified data to obtain her results. On June 4, 2014, Obokata agreed to retract the papers.
• Later (Investigation into disputed claims): In the months after the two Nature papers were released, all scientists who tried to duplicate Obokata's results failed and suspicion arose that her results were due to error or fraud.
• Later (Investigation into disputed claims): On April 1, RIKEN concluded that Obokata had engaged in "research misconduct", falsifying data on two occasions.
• Later (Investigation into disputed claims): On June 4, 2014, Obokata agreed to retract both the article and the "letter". The article was officially retracted on July 2, 2014.
• Later (Investigation into disputed claims): On September 24, 2015, the RIKEN scientists reported that Obokata's STAP cells came from embryonic stem cell contamination, while on the same day, research groups who had attempted to reproduce the STAP protocol jointly reported that they had found it irreproducible.
• Later (Implications): If the findings had proven to be valid, stimulus-triggered pluripotency cells could have been generated more easily and efficiently than by existing iPS techniques.

Other citations on it:

• NHS Choices (January 30, 2014). "Breakthrough in stem cell creation using acid bath - What did the research involve?". nhs.co.uk. Retrieved February 6, 2014. They put them in a weak acid solution (pH 5.7) for 30 minutes at 37°C, and then put them into petri dishes and grew them at normal pH.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-2081 / 41787889 / #2, read from revision 1375988571 of Stimulus-triggered acquisition of pluripotency.

The model's reasoning: The sentence presents the paper's claimed verification of pluripotency as established; contemporary news coverage plausibly repeats that finding.

Where the model says the article warns about this paper:

• In the opening paragraphs: Over the following months, all scientists who tried to duplicate her results failed, and suspicion arose that Obokata's results were due to error or fraud.
• In the opening paragraphs: On April 1, 2014, RIKEN concluded that Obokata had falsified data to obtain her results. On June 4, 2014, Obokata agreed to retract the papers.
• Later (Investigation into disputed claims): In the months after the two Nature papers were released, all scientists who tried to duplicate Obokata's results failed and suspicion arose that her results were due to error or fraud.
• Later (Investigation into disputed claims): On April 1, RIKEN concluded that Obokata had engaged in "research misconduct", falsifying data on two occasions.
• Later (Investigation into disputed claims): On June 4, 2014, Obokata agreed to retract both the article and the "letter". The article was officially retracted on July 2, 2014.
• Later (Investigation into disputed claims): On September 24, 2015, the RIKEN scientists reported that Obokata's STAP cells came from embryonic stem cell contamination, while on the same day, research groups who had attempted to reproduce the STAP protocol jointly reported that they had found it irreproducible.
• Later (Implications): If the findings had proven to be valid, stimulus-triggered pluripotency cells could have been generated more easily and efficiently than by existing iPS techniques.

Other citations on it:

• Thomson, Helen (January 29, 2014). "Stem cell power unleashed after 30 minute dip in acid". New Scientist. Retrieved January 31, 2014.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-2081 / 41787889 / #3, read from revision 1375988571 of Stimulus-triggered acquisition of pluripotency.

The model's reasoning: These survival and conversion measurements are results of the retracted study, presented without a local warning or another source.

Where the model says the article warns about this paper:

• In the opening paragraphs: Over the following months, all scientists who tried to duplicate her results failed, and suspicion arose that Obokata's results were due to error or fraud.
• In the opening paragraphs: On April 1, 2014, RIKEN concluded that Obokata had falsified data to obtain her results. On June 4, 2014, Obokata agreed to retract the papers.
• Later (Investigation into disputed claims): In the months after the two Nature papers were released, all scientists who tried to duplicate Obokata's results failed and suspicion arose that her results were due to error or fraud.
• Later (Investigation into disputed claims): On April 1, RIKEN concluded that Obokata had engaged in "research misconduct", falsifying data on two occasions.
• Later (Investigation into disputed claims): On June 4, 2014, Obokata agreed to retract both the article and the "letter". The article was officially retracted on July 2, 2014.
• Later (Investigation into disputed claims): On September 24, 2015, the RIKEN scientists reported that Obokata's STAP cells came from embryonic stem cell contamination, while on the same day, research groups who had attempted to reproduce the STAP protocol jointly reported that they had found it irreproducible.
• Later (Implications): If the findings had proven to be valid, stimulus-triggered pluripotency cells could have been generated more easily and efficiently than by existing iPS techniques.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-2081 / 41787889 / #4, read from revision 1375988571 of Stimulus-triggered acquisition of pluripotency.

The model's reasoning: The preceding uncited sentences and the footnoted sentence report the study's stress-response, chimera and germline-transmission findings without identifying them locally as retracted.

The sentence the footnote ends: Additionally, these mice produced healthy offspring, thereby demonstrating germline transmission which is "a strict criterion for pluripotency as well as genetic and epigenetic normality."

Where the model says the article warns about this paper:

• In the opening paragraphs: Over the following months, all scientists who tried to duplicate her results failed, and suspicion arose that Obokata's results were due to error or fraud.
• In the opening paragraphs: On April 1, 2014, RIKEN concluded that Obokata had falsified data to obtain her results. On June 4, 2014, Obokata agreed to retract the papers.
• Later (Investigation into disputed claims): In the months after the two Nature papers were released, all scientists who tried to duplicate Obokata's results failed and suspicion arose that her results were due to error or fraud.
• Later (Investigation into disputed claims): On April 1, RIKEN concluded that Obokata had engaged in "research misconduct", falsifying data on two occasions.
• Later (Investigation into disputed claims): On June 4, 2014, Obokata agreed to retract both the article and the "letter". The article was officially retracted on July 2, 2014.
• Later (Investigation into disputed claims): On September 24, 2015, the RIKEN scientists reported that Obokata's STAP cells came from embryonic stem cell contamination, while on the same day, research groups who had attempted to reproduce the STAP protocol jointly reported that they had found it irreproducible.
• Later (Implications): If the findings had proven to be valid, stimulus-triggered pluripotency cells could have been generated more easily and efficiently than by existing iPS techniques.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-2081 / 41787889 / #5, read from revision 1375988571 of Stimulus-triggered acquisition of pluripotency.

The model's reasoning: The proposed inhibitory mechanism interprets the paper's claimed reprogramming phenomenon; attribution and tentative language do not identify its retraction.

Where the model says the article warns about this paper:

• In the opening paragraphs: Over the following months, all scientists who tried to duplicate her results failed, and suspicion arose that Obokata's results were due to error or fraud.
• In the opening paragraphs: On April 1, 2014, RIKEN concluded that Obokata had falsified data to obtain her results. On June 4, 2014, Obokata agreed to retract the papers.
• Later (Investigation into disputed claims): In the months after the two Nature papers were released, all scientists who tried to duplicate Obokata's results failed and suspicion arose that her results were due to error or fraud.
• Later (Investigation into disputed claims): On April 1, RIKEN concluded that Obokata had engaged in "research misconduct", falsifying data on two occasions.
• Later (Investigation into disputed claims): On June 4, 2014, Obokata agreed to retract both the article and the "letter". The article was officially retracted on July 2, 2014.
• Later (Investigation into disputed claims): On September 24, 2015, the RIKEN scientists reported that Obokata's STAP cells came from embryonic stem cell contamination, while on the same day, research groups who had attempted to reproduce the STAP protocol jointly reported that they had found it irreproducible.
• Later (Implications): If the findings had proven to be valid, stimulus-triggered pluripotency cells could have been generated more easily and efficiently than by existing iPS techniques.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-2081 / 41787889 / #6, read from revision 1375988571 of Stimulus-triggered acquisition of pluripotency.

The model's reasoning: This historical account repeats the paper's claimed somatic-cell conversion as successful; unseen footnotes may supply historical reporting but require checking.

Where the model found it: The technique for producing STAP cells was subsequently studied by Obokata at the Brigham and Women's Hospital (BWH), while she was studying as a post doc under Charles Vacanti, and then at the RIKEN Center for Developmental Biology in Japan. In 2008, while working at Harvard Medical School, she verified at the request of Charles Vacanti that some of the cultured cells she was working with shrank to the size of stem cells after being mechanically injured in a capillary tube. She went on as directed, to test the effects of various stimuli on cells. After modifying the technique, Obokata was able to show that white blood cells from newborn mice could be transformed into cells that behaved much like stem cells. She repeated the experiment with other cell types including brain, skin, and muscle cells with the same result.

Where the model says the article warns about this paper:

• In the opening paragraphs: Over the following months, all scientists who tried to duplicate her results failed, and suspicion arose that Obokata's results were due to error or fraud.
• In the opening paragraphs: On April 1, 2014, RIKEN concluded that Obokata had falsified data to obtain her results. On June 4, 2014, Obokata agreed to retract the papers.
• In the same section (Investigation into disputed claims): In the months after the two Nature papers were released, all scientists who tried to duplicate Obokata's results failed and suspicion arose that her results were due to error or fraud.
• In the same section (Investigation into disputed claims): On April 1, RIKEN concluded that Obokata had engaged in "research misconduct", falsifying data on two occasions.
• In the same section (Investigation into disputed claims): On June 4, 2014, Obokata agreed to retract both the article and the "letter". The article was officially retracted on July 2, 2014.
• In the same section (Investigation into disputed claims): On September 24, 2015, the RIKEN scientists reported that Obokata's STAP cells came from embryonic stem cell contamination, while on the same day, research groups who had attempted to reproduce the STAP protocol jointly reported that they had found it irreproducible.
• Later (Implications): If the findings had proven to be valid, stimulus-triggered pluripotency cells could have been generated more easily and efficiently than by existing iPS techniques.

The model's confidence: 0.98. Finding RW-2081 / 41787889 / #8, read from revision 1375988571 of Stimulus-triggered acquisition of pluripotency.

The model's reasoning: These experiments retell the paper's claimed chimera contribution and lymphocyte reprogramming, without a local correction; an editor must check any sources attached to this history.

Where the model found it: Initially Obokata's findings were met with skepticism, even among her coworkers. "Everyone said it was an artefact – there were some really hard days", she recalled. The manuscript describing the work was rejected multiple times before its eventual publication as an article (together with a shorter jointly-written "letter") within the journal Nature. A series of experiments, first turning a mouse embryo green by fluorescently tagging STAP cells, then videotaping the transformation of T-cells into pluripotent cells, finally convinced skeptics that the results were real.

Where the model says the article warns about this paper:

• In the opening paragraphs: Over the following months, all scientists who tried to duplicate her results failed, and suspicion arose that Obokata's results were due to error or fraud.
• In the opening paragraphs: On April 1, 2014, RIKEN concluded that Obokata had falsified data to obtain her results. On June 4, 2014, Obokata agreed to retract the papers.
• In the same section (Investigation into disputed claims): In the months after the two Nature papers were released, all scientists who tried to duplicate Obokata's results failed and suspicion arose that her results were due to error or fraud.
• In the same section (Investigation into disputed claims): On April 1, RIKEN concluded that Obokata had engaged in "research misconduct", falsifying data on two occasions.
• In the same section (Investigation into disputed claims): On June 4, 2014, Obokata agreed to retract both the article and the "letter". The article was officially retracted on July 2, 2014.
• In the same section (Investigation into disputed claims): On September 24, 2015, the RIKEN scientists reported that Obokata's STAP cells came from embryonic stem cell contamination, while on the same day, research groups who had attempted to reproduce the STAP protocol jointly reported that they had found it irreproducible.
• Later (Implications): If the findings had proven to be valid, stimulus-triggered pluripotency cells could have been generated more easily and efficiently than by existing iPS techniques.

The model's confidence: 0.95. Finding RW-2081 / 41787889 / #9, read from revision 1375988571 of Stimulus-triggered acquisition of pluripotency.

The model's reasoning: The comparison uses the retracted study's claimed conversion efficiency as its benchmark; any separate source for the comparison must be checked.

Where the model found it: One previous way of creating stem cells has been via genetic manipulation of adult cells into iPS cells. Progress on iPS-based therapies has been slow due to regulatory hurdles surrounding genetic manipulation. Additionally, iPS techniques have an observed efficiency of around 1%, significantly lower than the claimed efficiency of STAP.

Where the model says the article warns about this paper:

• In the opening paragraphs: Over the following months, all scientists who tried to duplicate her results failed, and suspicion arose that Obokata's results were due to error or fraud.
• In the opening paragraphs: On April 1, 2014, RIKEN concluded that Obokata had falsified data to obtain her results. On June 4, 2014, Obokata agreed to retract the papers.
• Earlier (Investigation into disputed claims): In the months after the two Nature papers were released, all scientists who tried to duplicate Obokata's results failed and suspicion arose that her results were due to error or fraud.
• Earlier (Investigation into disputed claims): On April 1, RIKEN concluded that Obokata had engaged in "research misconduct", falsifying data on two occasions.
• Earlier (Investigation into disputed claims): On June 4, 2014, Obokata agreed to retract both the article and the "letter". The article was officially retracted on July 2, 2014.
• Earlier (Investigation into disputed claims): On September 24, 2015, the RIKEN scientists reported that Obokata's STAP cells came from embryonic stem cell contamination, while on the same day, research groups who had attempted to reproduce the STAP protocol jointly reported that they had found it irreproducible.
• In the same section (Implications): If the findings had proven to be valid, stimulus-triggered pluripotency cells could have been generated more easily and efficiently than by existing iPS techniques.

The model's confidence: 0.98. Finding RW-2081 / 41787889 / #10, read from revision 1375988571 of Stimulus-triggered acquisition of pluripotency.

The model's reasoning: Both sentences state the retracted trial's cardiovascular results as fact; the methodological and analytical errors undermine reliance on these estimates.

The sentence the footnote ends: The trial was planned for six years, but it was terminated early after a median follow-up of 4.8 years, and demonstrated that both Mediterranean diet groups reached a statistically significant reduction in the rate of the composite cardiovascular primary end-point of myocardial infarction, stroke, or cardiovascular death.

Where the model says the article warns about this paper:

• In the opening paragraphs: Parts of this article (those related to the retracted study need to be removed, with accompanying text, and replaced by the 2018 study) need to be updated.
• Later (Criticisms): The core publication was retracted and republished as a non-randomised study.
• Later (Network): The main paper with the final cardiovascular results was published in 2013, but retracted and immediately republished in 2018 because of errors in statistical methodology, which were revised in the republished paper.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-18405 / 52857727 / #1, read from revision 1344123348 of Predimed.

The model's reasoning: This uses PREDIMED's cardiovascular benefit as evidence without locally acknowledging its retraction; an editor should check whether this paragraph's own sources independently support the comparison.

Where the model found it: Some criticisms included that the protocol for the low-fat control group was changed during the trial, there were losses to follow up in the low-fat group which may have biased the results, and that a true low-fat diet was not achieved by the control group. Additionally, participants lived in a Mediterranean country. Thus, it is unclear if the results will be applicable to other non-Mediterranean settings. Despite these criticisms, there is no other dietary pattern with such a strong evidence of cardiovascular benefit, supported by this major trial and by the Lyon Diet-Heart Study but also reinforced by a multitude of other large and well-conducted observational studies. The transferability of the Mediterranean diet to other geographical regions far from the Mediterranean Sea represents the next main challenges.

Where the model says the article warns about this paper:

• In the opening paragraphs: Parts of this article (those related to the retracted study need to be removed, with accompanying text, and replaced by the 2018 study) need to be updated.
• In the same section (Criticisms): The core publication was retracted and republished as a non-randomised study.
• Later (Network): The main paper with the final cardiovascular results was published in 2013, but retracted and immediately republished in 2018 because of errors in statistical methodology, which were revised in the republished paper.

The model's confidence: 0.95. Finding RW-18405 / 52857727 / #3, read from revision 1344123348 of Predimed.

The model's reasoning: The caption presents the paper's reported helicase–DNA structure as established; the retraction for unreliable data and conclusions undermines that structural depiction.

The reference is tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-20239 / 53151252 / #1, read from revision 1315210353 of Large tumor antigen.

The model's reasoning: The pain-alleviation claim uses the review's finding as fact, although its healing claim and extension from Healing Touch, Therapeutic Touch and Reiki to consoling contact go beyond the abstract; the review was withdrawn as out of date.

Other citations on it:

• Meehan, Therese C. (July 1998). "Therapeutic touch as a nursing intervention: Integrative literature reviews and meta-analyses". Journal of Advanced Nursing. 28 (1): 117–125. doi:10.1046/j.1365-2648.1998.00771.x. PMID 9687138. S2CID 32762335.
• Whelan, Kathleen M.; Wishnia, Gracie S. (July 2003). "Reiki Therapy". Holistic Nursing Practice. 17 (4): 209–217. doi:10.1097/00004650-200307000-00008. ISSN 0887-9311. PMID 12889549. S2CID 34464263.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-32763 / 69259218 / #1, read from revision 1370316144 of Consoling touch.

The model's reasoning: With the list introduction, this asserts a TOP2B–BAZ1B (WSTF) interaction supported only by the retracted WINAC study; its data, image, and conclusion errors undermine that support.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.96. Finding RW-216 / 14138534 / #1, read from revision 1366113251 of TOP2B.

The model's reasoning: The paragraph presents specimen observations and interpretations from the withdrawn description as facts; its title directly supports the connection to the integument claim.

The sentence the footnote ends: It represents a juvenile, and possibly male, individual.

Where the model says the article warns about this paper:

• In the opening paragraphs: The taxon was informally named in 2020 in a now-withdrawn in-press academic paper.
• In the same section (Legal issues): Due to the ethical issues involving the potentially illegal transfer of the fossil from Brazil to Germany, the paper describing the specimen was "temporarily removed" only a few days after being made available online "in press" prior to formal publication. The article was later withdrawn in September 2021.
• In the same section (Nomenclatural issues): The "temporary removal" and ultimately the retraction of the publication describing the new genus and species, however, raise questions as to the nomenclatural validity of the new taxon.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.94. Finding RW-28669 / 66094894 / #1, read from revision 1360802178 of Ubirajara jubatus.

The model's reasoning: This specimen-specific size estimate appears to derive from the original description; an editor must check whether the paragraph has an independent source.

Where the model says the article warns about this paper:

• In the opening paragraphs: The taxon was informally named in 2020 in a now-withdrawn in-press academic paper.
• Earlier (Legal issues): Due to the ethical issues involving the potentially illegal transfer of the fossil from Brazil to Germany, the paper describing the specimen was "temporarily removed" only a few days after being made available online "in press" prior to formal publication. The article was later withdrawn in September 2021.
• Earlier (Nomenclatural issues): The "temporary removal" and ultimately the retraction of the publication describing the new genus and species, however, raise questions as to the nomenclatural validity of the new taxon.

The model's confidence: 0.83. Finding RW-28669 / 66094894 / #5, read from revision 1360802178 of Ubirajara jubatus.

The model's reasoning: This reports the original study's phylogenetic result; the accompanying criticism concerns reproducibility, not an explicit refutation or withdrawal warning, and an editor should check the paragraph's sources.

Where the model found it: The specimen was placed in the Compsognathidae family in 2020, as the sister species of a clade formed by Sinosauropteryx and Compsognathus. This phylogeny was criticized since it cannot be properly scientifically tested or replicated at the moment because the data supporting it were never made available.

Where the model says the article warns about this paper:

• In the opening paragraphs: The taxon was informally named in 2020 in a now-withdrawn in-press academic paper.
• Earlier (Legal issues): Due to the ethical issues involving the potentially illegal transfer of the fossil from Brazil to Germany, the paper describing the specimen was "temporarily removed" only a few days after being made available online "in press" prior to formal publication. The article was later withdrawn in September 2021.
• Earlier (Nomenclatural issues): The "temporary removal" and ultimately the retraction of the publication describing the new genus and species, however, raise questions as to the nomenclatural validity of the new taxon.

The model's confidence: 0.92. Finding RW-28669 / 66094894 / #6, read from revision 1360802178 of Ubirajara jubatus.

The model's reasoning: The caption describes the paper's problematic illustration and explicitly identifies its retraction, rather than relying on its scientific findings.

Where the model says the article warns about this paper:

• In the same section (In science): A now-retracted paper in Frontiers in Cell and Developmental Biology included this AI-generated rat with nonsensical anatomy and text.
• In the same section (In science): In 2024, an article containing a generated image of a rat with absurdly large genitals accompanied by nonsensical text and diagrams was retracted by the journal Frontiers in Cell and Developmental Biology after drawing attention from scientists on social media.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-51983 / 77986589 / #1, read from revision 1378426247 of AI slop.

The model's reasoning: The paper is cited as an example of AI-generated material in scientific publishing, not as evidence for its findings about stem cells or JAK/STAT signaling; the paragraph also explicitly describes its retraction.

Where the model says the article warns about this paper:

• In the same section (In science): A now-retracted paper in Frontiers in Cell and Developmental Biology included this AI-generated rat with nonsensical anatomy and text.
• In the same section (In science): In 2024, an article containing a generated image of a rat with absurdly large genitals accompanied by nonsensical text and diagrams was retracted by the journal Frontiers in Cell and Developmental Biology after drawing attention from scientists on social media.

Other citations on it:

• DeGeurin, Mack (19 March 2024). "AI-generated nonsense is leaking into scientific journals". Popular Science. Retrieved 5 July 2025.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-51983 / 77986589 / #2, read from revision 1378426247 of AI slop.

The model's reasoning: This particle-description claim overlaps the paper's experimental ultrastructure findings; an editor should check whether this paragraph has an independent source for the shape and dimensions.

Where the model found it: XMRV is a murine leukemia virus (MLV) that formed through the recombination of the genomes of two parent MLVs known as preXMRV-1 and preXMRV-2. MLVs belong to the virus family Retroviridae and the genus gammaretrovirus and have a single-stranded RNA genome that replicates through a DNA intermediate. The name XMRV was given because the discoverers of the virus initially thought that it was a novel potential human pathogen that was related to but distinct from MLVs. The XMRV particle is approximately spherical and 80 to 100 nm in diameter. Several XMRV genomic sequences have been published to date. These sequences are almost identical, an unusual finding as retroviruses replicate their genomes with relatively low fidelity, leading to divergent viral sequences in a single host organism. In 2010 the results of phylogenetic analyses of XMRV and related murine retroviruses led a group of researchers to conclude that XMRV "might not be a genuine human pathogen". Xenotropic viruses (xenos Gr. foreign; tropos Gr. turning) were initially discovered in the New Zealand Black (NZB) mouse and later found to be present in many other mouse strains including wild mice.

Where the model says the article warns about this paper:

• In the opening paragraphs: Initial reports erroneously linked the virus to prostate cancer and later to chronic fatigue syndrome (CFS)
• In the opening paragraphs: Subsequent research established that XMRV was in fact a laboratory contaminant, rather than a novel pathogen
• In the opening paragraphs: There is no evidence that XMRV infects humans, nor that XMRV is associated with or causes any human disease.
• Later (Prostate cancer): However, subsequent studies and retractions cast doubt on these findings.
• Later (Controversy and origins): Positive findings of the virus were likely due to contamination rather than true presence of the virus in humans.
• Later (Controversy and origins): As a result, many of the key publications which did claim an association were voluntarily retracted.

The model's confidence: 0.78. Finding RW-6343 / 9686427 / #3, read from revision 1360216464 of Xenotropic murine leukemia virus–related virus.

The model's reasoning: The asserted recovery from lymphocytes matches this paper's blood-cell findings and is used as evidence for transfusion concerns; an editor should check whether the paragraph has an independent source.

Where the model found it: Separate from these concerns, alarms were raised over the possibility that XMRV might be transmissible by blood transfusion since the virus was recovered from lymphocytes (white blood cells). XMRV is closely related to several known xenotropic mouse viruses which can recognize and enter cells of non-rodent species (including humans) by means of the cell surface xenotropic and polytropic retrovirus receptor 1 (XPR1). As a result, the AABB (formerly the American Association of Blood Banks) established a task force to determine the prevalence of XMRV in the United States' blood donation supply and the suitability of different detection methods.

Where the model says the article warns about this paper:

• In the opening paragraphs: Initial reports erroneously linked the virus to prostate cancer and later to chronic fatigue syndrome (CFS)
• In the opening paragraphs: Subsequent research established that XMRV was in fact a laboratory contaminant, rather than a novel pathogen, and had been generated unintentionally in the laboratory through genetic recombination between two mouse retroviruses during propagation of a prostate-cancer cell line in the mid-1990s.
• In the opening paragraphs: There is no evidence that XMRV infects humans, nor that XMRV is associated with or causes any human disease.
• In the same section (Controversy and origins): Positive findings of the virus were likely due to contamination rather than true presence of the virus in humans.
• In the same section (Controversy and origins): As a result, many of the key publications which did claim an association were voluntarily retracted. This included the initial study which had linked XMRV to CFS, which was retracted at Silverman's request; one of the co-authors, Judy Mikovits, was also accused of scientific misconduct.

The model's confidence: 0.93. Finding RW-492 / 9686427 / #3, read from revision 1360216464 of Xenotropic murine leukemia virus–related virus.

The model's reasoning: This retells the original paper's identification of a distinct potential human pathogen without a local correction; an editor should check whether the paragraph has an independent historical source.

Where the model found it: XMRV is a murine leukemia virus (MLV) that formed through the recombination of the genomes of two parent MLVs known as preXMRV-1 and preXMRV-2. MLVs belong to the virus family Retroviridae and the genus gammaretrovirus and have a single-stranded RNA genome that replicates through a DNA intermediate. The name XMRV was given because the discoverers of the virus initially thought that it was a novel potential human pathogen that was related to but distinct from MLVs. The XMRV particle is approximately spherical and 80 to 100 nm in diameter. Several XMRV genomic sequences have been published to date. These sequences are almost identical, an unusual finding as retroviruses replicate their genomes with relatively low fidelity, leading to divergent viral sequences in a single host organism. In 2010 the results of phylogenetic analyses of XMRV and related murine retroviruses led a group of researchers to conclude that XMRV "might not be a genuine human pathogen". Xenotropic viruses (xenos Gr. foreign; tropos Gr. turning) were initially discovered in the New Zealand Black (NZB) mouse and later found to be present in many other mouse strains including wild mice.

Where the model says the article warns about this paper:

• In the opening paragraphs: Initial reports erroneously linked the virus to prostate cancer and later to chronic fatigue syndrome (CFS)
• In the opening paragraphs: These findings raised serious questions concerning the findings of XMRV-related studies which purported to find connections between XMRV and human diseases. There is no evidence that XMRV infects humans, nor that XMRV is associated with or causes any human disease.
• Later (Prostate cancer): However, subsequent studies and retractions cast doubt on these findings.
• Later (Controversy and origins): Positive findings of the virus were likely due to contamination rather than true presence of the virus in humans.
• Later (Controversy and origins): As a result, many of the key publications which did claim an association were voluntarily retracted.

The model's confidence: 0.80. Finding RW-2195 / 9686427 / #4, read from revision 1360216464 of Xenotropic murine leukemia virus–related virus.

The model's reasoning: The discovery narrative repeats the paper's reported detection in prostate cancer samples without a local caveat; an editor should check any sources attached to this paragraph.

Where the model found it: XMRV was discovered in the laboratories of Joseph DeRisi at the University of California, San Francisco, and Robert Silverman and Eric Klein of the Cleveland Clinic. Silverman had previously cloned and investigated the enzyme ribonuclease L (RNase L), part of the cell's natural defense against viruses. When activated, RNase L degrades cellular and viral RNA to halt viral replication. In 2002, the "hereditary prostate cancer 1" locus (HPC1) was mapped to the RNase L gene, implicating it in the development of prostate cancer. The cancer-associated "R462Q" mutation results in a glutamine instead of an arginine at position 462 of the RNase L enzyme, reducing its catalytic activity. A man with two copies of this mutation has twice the risk of prostate cancer; one copy raises the risk by 50%. Klein and Silverman hypothesized that "the putative linkage of RNase L alterations to HPC might reflect enhanced susceptibility to a viral agent" and conducted a viral screen of prostate cancer samples, leading to the discovery of XMRV.

Where the model says the article warns about this paper:

• In the opening paragraphs: Initial reports erroneously linked the virus to prostate cancer and later to chronic fatigue syndrome (CFS)
• In the opening paragraphs: These findings raised serious questions concerning the findings of XMRV-related studies which purported to find connections between XMRV and human diseases. There is no evidence that XMRV infects humans, nor that XMRV is associated with or causes any human disease.
• Later (Prostate cancer): However, subsequent studies and retractions cast doubt on these findings.
• Later (Controversy and origins): Positive findings of the virus were likely due to contamination rather than true presence of the virus in humans.
• Later (Controversy and origins): As a result, many of the key publications which did claim an association were voluntarily retracted.

The model's confidence: 0.87. Finding RW-2195 / 9686427 / #5, read from revision 1360216464 of Xenotropic murine leukemia virus–related virus.

The model's reasoning: This reports the withdrawn review's efficacy conclusion without a local warning; the abstract identifies the pooled trials as TMD trials, not burning mouth syndrome trials.

Where the model says the article warns about this paper:

• Later (Prognosis): The article cited as the source of this information was withdrawn from publication, saying the information was out of date and not meeting Cochrane methodological standards.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-31688 / 33398582 / #5, read from revision 1314216559 of Atypical facial pain.

The model's reasoning: The sentence primarily supplies approval history, although 'improved' could imply comparative efficacy; the missing abstract leaves that connection uncertain, and an independent source is needed.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.76. Finding RW-32710 / 11265221 / #3, read from revision 1373833464 of 5-HT3 antagonist.

The model's reasoning: The paragraph attributes the arsenate-incorporation finding to the study without saying it was refuted or retracted; criticism alone does not provide that notice.

Where the model says the article warns about this paper:

• In the opening paragraphs: In 2012, two reports refuting the most significant aspects of the original results were published in the same journal in which the original findings had been previously published.
• In the opening paragraphs: On July 24, 2025, Science formally retracted the 2010 paper, citing expanded retraction criteria and concluding that the experiments did not support the paper's key conclusions; the study's authors objected to the decision.
• In the same section (GFAJ-1 controversy): Following the publication of the articles challenging the conclusions of the original Science article first describing GFAJ-1, David Sanders on the website Retraction Watch argued that the original article should be retracted because of misrepresentation of critical data.
• In the same section (GFAJ-1 controversy): In October 2024, Science editor Holden Thorp notified the article's authors of its intention to retract, arguing that, whereas formerly only misconduct justified retraction, current practice allows it for unreliability. The paper was retracted in 2025 with "no evidence of misconduct by researchers".

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.97. Finding RW-66335 / 29837208 / #2, read from revision 1372843758 of Felisa Wolfe-Simon.

The model's reasoning: The external-link title repeats the study's arsenic-based-life claim without a local refutation notice; an editor should check whether the linked press conference is sufficient historical context.

Where the model found it: NASA-funded research discovers life built with toxic chemical one-hour press conference on NASA TV 2010-12-02, with Wolfe-Simon, Mary Voytek, Steven A. Benner, Pamela Conrad and James Elser

Where the model says the article warns about this paper:

• In the opening paragraphs: In 2012, two reports refuting the most significant aspects of the original results were published in the same journal in which the original findings had been previously published.
• In the opening paragraphs: On July 24, 2025, Science formally retracted the 2010 paper, citing expanded retraction criteria and concluding that the experiments did not support the paper's key conclusions; the study's authors objected to the decision.
• Earlier (GFAJ-1 controversy): Following the publication of the articles challenging the conclusions of the original Science article first describing GFAJ-1, David Sanders on the website Retraction Watch argued that the original article should be retracted because of misrepresentation of critical data.
• Earlier (GFAJ-1 controversy): In October 2024, Science editor Holden Thorp notified the article's authors of its intention to retract, arguing that, whereas formerly only misconduct justified retraction, current practice allows it for unreliability. The paper was retracted in 2025 with "no evidence of misconduct by researchers".

The model's confidence: 0.80. Finding RW-66335 / 29837208 / #4, read from revision 1372843758 of Felisa Wolfe-Simon.

The model's reasoning: Both sentences present the paper’s reported wheat growth and oxidative-stress benefits as fact; the second continues the finding beyond the footnote.

The sentence the footnote ends: The application of BR successfully alleviate drought stress and improve wheat growth under deficit irrigation system.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-72297 / 670855 / #1, read from revision 1325494408 of Brassinosteroid.

The model's reasoning: These orthopedic examples rely on the paper's reported detection of periodontal pathogens in knee arthroplasty tissues and osteoarthritis aspirates; the other citations concern shunts and wounds, not these orthopedic findings.

The sentence the footnote ends: In particular, he has used these approaches in the study of orthopedic, neurosurgical, and wound infections to identify both known and novel pathogens in multiple conditions including: periprosthetic 'sterile' loosenings; suspected infected arthroplasties; 'aseptic' arthritis; and ventroperitoneal shunts as well as chronic non-healing wounds.

Other citations on it:

• Fux, C. A.; Quigley, M.; Worel, A. M.; Post, C.; Zimmerli, S.; Ehrlich, G.; Veeh, R. H. (April 4, 2006). "Biofilm-related infections of cerebrospinal fluid shunts". Clinical Microbiology and Infection. 12 (4): 331–337. doi:10.1111/j.1469-0691.2006.01361.x. PMID 16524409.
• Tuttle, M. S.; Mostow, E.; Mukherjee, P.; Hu, F. Z.; Melton-Kreft, R.; Ehrlich, G. D.; Dowd, S. E.; Ghannoum, M. A. (2011). "Characterization of Bacterial Communities in Venous Insufficiency Wounds by Use of Conventional Culture and Molecular Diagnostic Methods - PMC". Journal of Clinical Microbiology. 49 (11): 3812–3819. doi:10.1128/JCM.00847-11. PMC 3209111. PMID 21880958.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.91. Finding RW-14047 / 72478180 / #1, read from revision 1370418144 of Garth Ehrlich.

The model's reasoning: The resident cardiac stem-cell differentiation claim matches the paper's central finding, and the following sentence uses that claimed regenerative capacity to justify a therapeutic approach.

The sentence the footnote ends: Mammal hearts also host naturally occurring cardiac stem cells which may be capable of differentiating themselves into cardiomyocytes, endothelial cells and cardiac fibroblasts.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-19655 / 7024350 / #1, read from revision 1371854606 of Cardiomyoplasty.

The model's reasoning: This contraindication overlaps the paper's stated exclusion of patients with previous donor-site surgery; the abstract does not establish the other listed contraindications.

The sentence the footnote ends: The woman is not a suitable patient for a free-flap breast reconstruction surgical procedure if she has any of the following symptoms, or a combination of these symptoms: ASA III or an ASA IV surgical-health grade, a blood coagulation disorder, an unstable psychiatric disease, a BMI > 35 obesity grade, a previous surgery that interrupted the blood supply to the DIEAP free flap, or contraindications to anticoagulation therapy.

Other citations on it:

• Ahmed S.; Snelling A.; Bains M.; et al. (2005). "Breast reconstruction, clinical review". BMJ. 330 (7497): 943–8. doi:10.1136/bmj.330.7497.943. PMC 556341. PMID 15845976.
• Chevray Pierre M (May 2004). "Update on breast reconstruction using free TRAM, DIEP, and SIEA flaps". Semin Plast Surg. 18 (2): 97–104. doi:10.1055/s-2004-829044. PMC 2884730. PMID 20574488.
• Tseng Charles Y.; Lipa Joan E. (2010). "Perforator flaps in breast reconstruction". Clin Plastic Surg. 37 (4): 641–654. doi:10.1016/j.cps.2010.06.002. PMID 20816519.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.91. Finding RW-2516 / 37549699 / #1, read from revision 1372402828 of Free flap breast reconstruction.

The model's reasoning: The cross-reference imports the DIEAP contraindications, including previous surgery affecting blood supply; the additional artery-specific exclusion is not established by the abstract.

Other citations on it:

• Chevray Pierre M (May 2004). "Update on breast reconstruction using free TRAM, DIEP, and SIEA flaps". Semin Plast Surg. 18 (2): 97–104. doi:10.1055/s-2004-829044. PMC 2884730. PMID 20574488.
• Tseng Charles Y.; Lipa Joan E. (2010). "Perforator flaps in breast reconstruction". Clin Plastic Surg. 37 (4): 641–654. doi:10.1016/j.cps.2010.06.002. PMID 20816519.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.85. Finding RW-2516 / 37549699 / #2, read from revision 1372402828 of Free flap breast reconstruction.

The model's reasoning: This states the paper's donor-site surgery and liposuction contraindication as fact; other cited flap reviews plausibly support it, and the retraction concerns duplication.

Other citations on it:

• Tseng Charles Y.; Lipa Joan E. (2010). "Perforator flaps in breast reconstruction". Clin Plastic Surg. 37 (4): 641–654. doi:10.1016/j.cps.2010.06.002. PMID 20816519.
• Guerra AB, Soueid N, Metzinger SE, Levine J, Bidros RS, Erhard H, Allen RJ. Simultaneous Bilateral Breast Reconstruction with Superior Gluteal Artery Perforator (SGAP) flaps. Annals of Plastic Surgery. 2004 Oct;53(4):305-10
• Granzow JW, Levine JL, Chiu ES, Allen RJ (2006). "Breast Reconstruction with Gluteal Artery Perforator Flaps". Journal of Plastic, Reconstructive & Aesthetic Surgery. 59 (6): 614–21. doi:10.1016/j.bjps.2006.01.005. PMID 16716955.
• Rad AN, Flores JI, Prucz RB, Stapleton SM, Rosson GD. Clinical Experience with the Lateral Septocutaneous Superior Gluteal Artery Perforator Flap for Autologous Breast Reconstruction. Microsurgery. 2010 July; 30(5):339-47
• Baumeister S, Werdin F, Peek A (May 2010). "The SGAP flap: Rare Exception or Second choice in Autologous Breast Reconstruction?". Journal of Reconstructive Microsurgery. 26 (4): 251–8. doi:10.1055/s-0030-1248233. PMID 20169527.
• LoTempio MM, Allen RJ (August 2010). "Breast Reconstruction with SGAP and IGAP flaps". Plastic and Reconstructive Surgery. 126 (2): 393–401. doi:10.1097/prs.0b013e3181de236a. PMID 20679825. S2CID 20950415.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.98. Finding RW-2516 / 37549699 / #3, read from revision 1372402828 of Free flap breast reconstruction.

The model's reasoning: Under Contraindication, this imports the SGAP exclusion for previous surgery such as liposuction, matching the paper's stated contraindications.

Other citations on it:

• Tseng Charles Y.; Lipa Joan E. (2010). "Perforator flaps in breast reconstruction". Clin Plastic Surg. 37 (4): 641–654. doi:10.1016/j.cps.2010.06.002. PMID 20816519.
• Granzow JW, Levine JL, Chiu ES, Allen RJ (2006). "Breast Reconstruction with Gluteal Artery Perforator Flaps". Journal of Plastic, Reconstructive & Aesthetic Surgery. 59 (6): 614–21. doi:10.1016/j.bjps.2006.01.005. PMID 16716955.
• LoTempio MM, Allen RJ (August 2010). "Breast Reconstruction with SGAP and IGAP flaps". Plastic and Reconstructive Surgery. 126 (2): 393–401. doi:10.1097/prs.0b013e3181de236a. PMID 20679825. S2CID 20950415.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-2516 / 37549699 / #4, read from revision 1372402828 of Free flap breast reconstruction.

The model's reasoning: Under MS-TRAM Contraindications, this imports the DIEAP exclusions supported by the retracted paper, including previous surgery affecting blood supply; an editor should check whether this item has independent sourcing.

The model's confidence: 0.83. Finding RW-2516 / 37549699 / #5, read from revision 1372402828 of Free flap breast reconstruction.

The model's reasoning: Under Septocutaneous GAP Contraindications, this repeats the SGAP surgery and liposuction exclusion supported by the retracted paper; an editor should check for independent sourcing of this item.

The model's confidence: 0.94. Finding RW-2516 / 37549699 / #6, read from revision 1372402828 of Free flap breast reconstruction.

The model's reasoning: This sentence reproduces the review's assessment of uncertain comparative outcomes; the later jug-use sentence is not a finding of the review.

The sentence the footnote ends: A jug may also be used by wheelchair users, rather than struggling to get over a toilet.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-30915 / 32017656 / #1, read from revision 1340894376 of Intermittent catheterisation.

The model's reasoning: Reports the paper's central superconductivity claim without identifying its retraction; skepticism and failed replication do not explicitly establish that this claim was refuted.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-48932 / 63919070 / #1, read from revision 1351313933 of Hydridonitride.

The model's reasoning: Presents the reported material's composition, structure, pressure range and colour as established properties; the accompanying Nature commentary plausibly discusses the same observations.

Other citations on it:

• Jin, ChangQing; Ceperly, David (8 March 2023). "Hopes raised for room-temperature superconductivity, but doubts remain". Nature. 615 (7951): 221–222. Bibcode:2023Natur.615..221J. doi:10.1038/d41586-023-00599-9. PMID 36890377. S2CID 257407330.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.94. Finding RW-48932 / 63919070 / #3, read from revision 1351313933 of Hydridonitride.

The model's reasoning: States the retracted superconductivity result and associated material properties as fact, with no other source or local retraction notice.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-48932 / 63919070 / #4, read from revision 1351313933 of Hydridonitride.

The model's reasoning: This reports the retracted trial's efficacy and bone-density claims without a local warning; attribution to the study does not remove reliance on its findings.

Where the model says the article warns about this paper:

• In the opening paragraphs: Jamal published a high-profile paper suggesting that the heart medication nitroglycerin was a treatment for osteoporosis, and was later demonstrated to have misrepresented her results.
• Later (Misconduct controversy): When Eastell conducted his own statistical analysis in 2014, he found no difference between the treatment and placebo groups.
• Later (Misconduct controversy): Following investigation by the Women's College Hospital, Jamal was deemed to have manipulated data and misrepresented her findings.
• Later (Misconduct controversy): That December, her paper on nitroglycerin was retracted by JAMA, the journal of the American Medical Association and its original publisher.
• Later (Misconduct controversy): Following the JAMA retraction, two further papers of Jamal's were retracted; one on the risk of osteoporosis in kidney disease, and one on nitrate use and bone density.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-4478 / 66974230 / #1, read from revision 1375196474 of Sophie Jamal.

The model's reasoning: The p53 claim matches the paper's reported reduction in p53 expression in infected mice; the following pleural-effusion sentence is not a finding established by the supplied abstract.

The sentence the footnote ends: There have also been cases of malignant pleural effusion, when patients exhibited chronic M. penetrans infection with various immunodeficiencies (such as HIV infections or anticancer treatment).

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-24457 / 40234896 / #1, read from revision 1354661092 of Malacoplasma penetrans.

The model's reasoning: The editorial is reporting this paper's neurotoxicity finding and risk inference; an editor should check any local sources because attribution alone does not independently establish the result.

Where the model says the article warns about this paper:

• In the opening paragraphs: It was later retracted; instead of using MDMA, methamphetamine had been used in the test.
• Later (Retraction): Retraction
• Later (Retraction): In September 2003, the paper was retracted. In a statement published in Science, the research team indicated that due to a labelling error, methamphetamine had been administered to 9 of the 10 test animals instead of MDMA. The team had consistently been unable to replicate the original results, which lead to them conducting an investigation and ultimately discovering the error.
• Later (Retraction): Following the retraction, Ricaurte stated that he would continue to investigate the possibility of a relationship between MDMA and dopamine dysfunction, and that the laboratory would be adjusting its chemical handling procedure.
• Later (Reactions): We shared in one: Some vials containing the recreational drug Ecstasy got switched with vials containing methamphetamine, and we wound up publishing a paper we wish we hadn't
• Later (Reactions): Journalists such as Larry Smith and Carla Spartos have stated that the inaccurate study may have influenced drug policy being made at the time, such as the RAVE act of April 2003.
• Later (Reactions): A study of the 50 largest American newspapers with available online archives found that 26 (52%) covered the original article's publication, while 20 (40%) covered its retraction. Of the newspapers that covered the article's publication, only 14 (54%) covered its retraction, suggesting large numbers of newspaper readers who had been exposed to the misinformation were never made aware that the study had been in error.

The model's confidence: 0.98. Finding RW-961 / 653532 / #2, read from revision 1370020709 of Retracted article on dopaminergic neurotoxicity of MDMA.

The model's reasoning: This reports the harm inference drawn from the retracted experiment through news coverage; an editor should check the paragraph's sources, which are not visible here.

Where the model found it: Following the release of the paper, Science published a “News of the week” article by Constance Holden. The article noted that the results of the study had caused the researchers concern that even a single night of MDMA usage could cause brain damage, and leave a person vulnerable to neurological disorders such as Parkinson's disease. Holden also noted that the findings of the study were surprising, due to MDMA being known for prompting the release of large amounts of serotonin, but not dopamine. In the article, cognitive neuroscientist Jon Cole was described as being “skeptical” about the risk of Parkinson's from MDMA use, stating that there had only been one case report of Parkinson's related to the use of ecstasy. In response, the researchers stated that this could be due to symptoms not presenting until "70% to 80%" of dopamine had been depleted.

Where the model says the article warns about this paper:

• In the opening paragraphs: It was later retracted; instead of using MDMA, methamphetamine had been used in the test.
• Later (Retraction): Retraction
• Later (Retraction): In September 2003, the paper was retracted. In a statement published in Science, the research team indicated that due to a labelling error, methamphetamine had been administered to 9 of the 10 test animals instead of MDMA. The team had consistently been unable to replicate the original results, which lead to them conducting an investigation and ultimately discovering the error.
• Later (Retraction): Following the retraction, Ricaurte stated that he would continue to investigate the possibility of a relationship between MDMA and dopamine dysfunction, and that the laboratory would be adjusting its chemical handling procedure.
• Later (Reactions): We shared in one: Some vials containing the recreational drug Ecstasy got switched with vials containing methamphetamine, and we wound up publishing a paper we wish we hadn't
• Later (Reactions): Journalists such as Larry Smith and Carla Spartos have stated that the inaccurate study may have influenced drug policy being made at the time, such as the RAVE act of April 2003.
• Later (Reactions): A study of the 50 largest American newspapers with available online archives found that 26 (52%) covered the original article's publication, while 20 (40%) covered its retraction. Of the newspapers that covered the article's publication, only 14 (54%) covered its retraction, suggesting large numbers of newspaper readers who had been exposed to the misinformation were never made aware that the study had been in error.

The model's confidence: 0.97. Finding RW-961 / 653532 / #3, read from revision 1370020709 of Retracted article on dopaminergic neurotoxicity of MDMA.

The model's reasoning: The quotation uses this study as evidence of acute risk rather than supplying independent evidence; an editor should verify its local sourcing and historical context.

Where the model says the article warns about this paper:

• In the opening paragraphs: It was later retracted; instead of using MDMA, methamphetamine had been used in the test.
• Later (Retraction): Retraction
• Later (Retraction): In September 2003, the paper was retracted. In a statement published in Science, the research team indicated that due to a labelling error, methamphetamine had been administered to 9 of the 10 test animals instead of MDMA. The team had consistently been unable to replicate the original results, which lead to them conducting an investigation and ultimately discovering the error.
• Later (Retraction): Following the retraction, Ricaurte stated that he would continue to investigate the possibility of a relationship between MDMA and dopamine dysfunction, and that the laboratory would be adjusting its chemical handling procedure.
• Later (Reactions): We shared in one: Some vials containing the recreational drug Ecstasy got switched with vials containing methamphetamine, and we wound up publishing a paper we wish we hadn't
• Later (Reactions): Journalists such as Larry Smith and Carla Spartos have stated that the inaccurate study may have influenced drug policy being made at the time, such as the RAVE act of April 2003.
• Later (Reactions): A study of the 50 largest American newspapers with available online archives found that 26 (52%) covered the original article's publication, while 20 (40%) covered its retraction. Of the newspapers that covered the article's publication, only 14 (54%) covered its retraction, suggesting large numbers of newspaper readers who had been exposed to the misinformation were never made aware that the study had been in error.

The model's confidence: 0.94. Finding RW-961 / 653532 / #4, read from revision 1370020709 of Retracted article on dopaminergic neurotoxicity of MDMA.

The model's reasoning: This linked news title repeats the distinctive neurotoxicity claim of the paper; an editor should check the linked source to confirm its dependence and how the historical link is presented.

Where the model found it: Trent S (30 September 2002). Perschetz L (ed.). "Recreational Use of Ecstasy Causes New Brain Damage: Trend to sequential doses of popular drug can have long-term lasting effects". The JHU Gazette (Johns Hopkins University Gazette). 32 (5). Baltimore, Maryland, United States of America: Johns Hopkins University. Archived from the original on 21 April 2017. Retrieved 10 July 2021.

Where the model says the article warns about this paper:

• In the opening paragraphs: It was later retracted; instead of using MDMA, methamphetamine had been used in the test.
• Earlier (Retraction): Retraction
• Earlier (Retraction): In September 2003, the paper was retracted. In a statement published in Science, the research team indicated that due to a labelling error, methamphetamine had been administered to 9 of the 10 test animals instead of MDMA. The team had consistently been unable to replicate the original results, which lead to them conducting an investigation and ultimately discovering the error.
• Earlier (Retraction): Following the retraction, Ricaurte stated that he would continue to investigate the possibility of a relationship between MDMA and dopamine dysfunction, and that the laboratory would be adjusting its chemical handling procedure.
• Earlier (Reactions): We shared in one: Some vials containing the recreational drug Ecstasy got switched with vials containing methamphetamine, and we wound up publishing a paper we wish we hadn't
• Earlier (Reactions): Journalists such as Larry Smith and Carla Spartos have stated that the inaccurate study may have influenced drug policy being made at the time, such as the RAVE act of April 2003.
• Earlier (Reactions): A study of the 50 largest American newspapers with available online archives found that 26 (52%) covered the original article's publication, while 20 (40%) covered its retraction. Of the newspapers that covered the article's publication, only 14 (54%) covered its retraction, suggesting large numbers of newspaper readers who had been exposed to the misinformation were never made aware that the study had been in error.

The model's confidence: 0.86. Finding RW-961 / 653532 / #5, read from revision 1370020709 of Retracted article on dopaminergic neurotoxicity of MDMA.

The model's reasoning: This reports the withdrawn review's quantitative treatment-response finding without qualification; withdrawal as out of date calls for a current source rather than automatic removal of the claim.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-31031 / 25045060 / #1, read from revision 1365076664 of Prevention of migraine attacks.

The model's reasoning: This reports the withdrawn review's pooled efficacy finding; the review was withdrawn as out of date and replaced, rather than invalidated for misconduct.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-31392 / 25045060 / #1, read from revision 1365076664 of Prevention of migraine attacks.

The model's reasoning: This introductory paragraph repeats the class-wide efficacy conclusion of the withdrawn review cited immediately afterward; an editor should check whether it has an independent source.

The model's confidence: 0.85. Finding RW-31392 / 25045060 / #2, read from revision 1365076664 of Prevention of migraine attacks.

The model's reasoning: The title establishes the lunglessness claim but does not show whether the locality count was a finding of this paper or background information.

Where the model says the article warns about this paper:

• In the opening paragraphs: However, micro CT scanning revealed that, like all other known frog species, the Bornean flat-headed frog has lungs, though they are tiny.
• In the opening paragraphs: Bickford et al. dissected some of the specimens and were unable to see the lungs, leading them to incorrectly conclude that no lungs were present.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.93. Finding RW-55499 / 12358044 / #3, read from revision 1370514695 of Bornean flat-headed frog.

The model's reasoning: The external-link title repeats the paper's distinctive lunglessness finding without a correction; an editor should check whether the linked report covers this study and has been updated.

Where the model says the article warns about this paper:

• In the opening paragraphs: However, micro CT scanning revealed that, like all other known frog species, the Bornean flat-headed frog has lungs, though they are tiny.
• In the opening paragraphs: Bickford et al. dissected some of the specimens and were unable to see the lungs, leading them to incorrectly conclude that no lungs were present.

The model's confidence: 0.88. Finding RW-55499 / 12358044 / #5, read from revision 1370514695 of Bornean flat-headed frog.

The model's reasoning: This reports the retracted paper's central finding without a retraction warning in this paragraph; the legal grounds for retraction do not establish that the finding is invalid.

Where the model says the article warns about this paper:

• In the same section ("Recursive fury"): The paper was retracted with a notice published in March 2014, which stated:
• In the same section ("Recursive fury"): It did, however, determine that the legal context is insufficiently clear and therefore Frontiers wishes to retract the published article.
• In the same section ("Recursive fury"): Frontiers is concerned about solid science, and it's obviously a regret when you have to retract an article that is scientifically and ethically sound.
• In the same section ("Recursive fury"): said that she was "profoundly disappointed" by its retraction
• In the same section ("Recursive fury"): The Sydney Morning Herald reported that academics had described the retraction as having "a chilling effect on research".
• In the same section ("Recursive fury"): After media reports about this retraction, the journal's Editorial Director Costanza Zucca and Executive Editor Fred Fenter made a joint statement on the journal's blog that "Frontiers did not 'cave in to threats'; in fact, Frontiers received no threats."
• In the same section ("Recursive fury"): The statement said that the main reason for retraction was insufficient protection for the rights of the studied subjects.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-4849 / 41915080 / #1, read from revision 1372462597 of Stephan Lewandowsky.

The model's reasoning: These attributed findings rely on the retracted study; the correction is explained only in the following paragraph, not in this paragraph.

The sentence the footnote ends: The authors reported that children from religious households believe that interpersonal harm is more "mean" and deserving of harsher punishment than non-religious children.

Where the model says the article warns about this paper:

• In the same section (Controversies): However, Decety has since retracted the study, citing an analysis error that nullified the study's conclusion on religiosity and altruism.
• In the same section (Controversies): This came about after Azim F. Shariff reanalyzed the original data and observed that there was a mistake in the original study: coding country as a continuous rather than categorical variable.
• In the same section (Controversies): Once this error was corrected, as Shariff wrote, "most of the associations they observed with religious affiliation appear to be artifacts of between-country differences, driven primarily by low levels of generosity in Turkey and South Africa. However, children from highly religious households do appear slightly less generous than those from moderately religious ones."
• In the same section (Controversies): Decety has stated "When we reanalyzed these data to correct this error, we found that country of origin, rather than religious affiliation, is the primary predictor of several of the outcomes."

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-21002 / 8738925 / #1, read from revision 1375515692 of Jean Decety.

The model's reasoning: The citation's paragraph presents the paper's purported randomisation and response results as facts, without a local warning about fabrication.

The sentence the footnote ends: Bezwoda published the first randomised controlled trial of high-dose versus conventional-dose chemotherapy as first-line treatment for metastatic breast cancer in October 1995.

Where the model says the article warns about this paper:

• Earlier (History): but his study was later discredited due to scientific misconduct, including the falsification of data.
• Earlier (Funding mechanisms and litigation): although his study was later confirmed to have falsified data.
• In the same section (Randomised controlled trials): However, Bezwoda's research work was seriously flawed, and an audit concluded that the results were essentially fabricated.
• In the same section (Research misconduct): Research misconduct
• In the same section (Research misconduct): Treatment details bore little resemblance to the published data.
• In the same section (Research misconduct): The Journal of Clinical Oncology retracted the 1995 paper on April 26, 2001, concluding: "the trial was not conducted in a scientifically acceptable manner. The protocol was apparently written 9 years after the study was started and only after another study by the same investigator was to be audited. No patient signed a consent form, and there is little evidence of true randomisation." ASCO President Lawrence H. Einhorn called the fraud "a stunning betrayal of public trust".
• Later (Status in breast cancer): the finding that the Bezwoda study was false

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-3572 / 38028125 / #1, read from revision 1346314034 of High-dose chemotherapy and bone marrow transplant.

The model's reasoning: These response figures come directly from the fabricated trial and are presented as factual results in a historical account of the study.

Where the model says the article warns about this paper:

• Earlier (History): but his study was later discredited due to scientific misconduct, including the falsification of data.
• Earlier (Funding mechanisms and litigation): although his study was later confirmed to have falsified data.
• In the same section (Randomised controlled trials): However, Bezwoda's research work was seriously flawed, and an audit concluded that the results were essentially fabricated.
• In the same section (Research misconduct): Research misconduct
• In the same section (Research misconduct): Treatment details bore little resemblance to the published data.
• In the same section (Research misconduct): The Journal of Clinical Oncology retracted the 1995 paper on April 26, 2001, concluding: "the trial was not conducted in a scientifically acceptable manner. The protocol was apparently written 9 years after the study was started and only after another study by the same investigator was to be audited. No patient signed a consent form, and there is little evidence of true randomisation." ASCO President Lawrence H. Einhorn called the fraud "a stunning betrayal of public trust".
• Later (Status in breast cancer): the finding that the Bezwoda study was false

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-3572 / 38028125 / #2, read from revision 1346314034 of High-dose chemotherapy and bone marrow transplant.

The model's reasoning: The preceding description appears to concern preliminary results of the trial published in 1995, but an editor must check the paragraph's sources to confirm that connection.

Where the model found it: In May 1992, researchers from the University of the Witwatersrand (Wits) in South Africa presented early results of a randomised control trial (RCT) at the annual meeting of the American Society of Clinical Oncology (ASCO) in San Diego. The lead investigator was Werner Bezwoda, professor of haematology and oncology at Wits. This was the first scientific evidence of clinical benefit to emerge.

Where the model says the article warns about this paper:

• Earlier (History): but his study was later discredited due to scientific misconduct, including the falsification of data.
• Earlier (Funding mechanisms and litigation): although his study was later confirmed to have falsified data.
• In the same section (Randomised controlled trials): However, Bezwoda's research work was seriously flawed, and an audit concluded that the results were essentially fabricated.
• In the same section (Research misconduct): Research misconduct
• In the same section (Research misconduct): Treatment details bore little resemblance to the published data.
• In the same section (Research misconduct): The Journal of Clinical Oncology retracted the 1995 paper on April 26, 2001, concluding: "the trial was not conducted in a scientifically acceptable manner. The protocol was apparently written 9 years after the study was started and only after another study by the same investigator was to be audited. No patient signed a consent form, and there is little evidence of true randomisation." ASCO President Lawrence H. Einhorn called the fraud "a stunning betrayal of public trust".
• Later (Status in breast cancer): the finding that the Bezwoda study was false

The model's confidence: 0.86. Finding RW-3572 / 38028125 / #5, read from revision 1346314034 of High-dose chemotherapy and bone marrow transplant.

The model's reasoning: The serotype-specific contrast repeats the study's cell-culture finding that only APMV-1, -3 and -5 induced syncytia, without acknowledging retraction.

The sentence the footnote ends: In cell culture, many paramyxoviruses express a cytopathic effect by syncytium formation, the fusion of infected cells with neighboring cells producing a large multi-nucleated cell.

The reference is tagged {{Retracted}}.

The model's confidence: 0.93. Finding RW-24834 / 55694123 / #7, read from revision 1361943141 of Avian metaavulavirus 2.

The model's reasoning: This appears to repeat the study's respiratory-tract tropism finding; an editor should check whether this paragraph has an independent source.

Where the model found it: Avian metaavulavirus 2 is believed to infect tissues of the respiratory tract and in some cases, tissues of the lungs. The viral envelope protein, Hn, binds to sialic acid residues present in cells in the upper respiratory tract. It has been found that in chickens, some may present mild respiratory symptoms or be asymptomatic. It is believed that more severe symptoms including sinusitis and a decrease in egg production occurs in turkeys. Other symptoms including reduced hatchability and infertility of eggs has been reported.

The model's confidence: 0.77. Finding RW-24834 / 55694123 / #8, read from revision 1361943141 of Avian metaavulavirus 2.

The model's reasoning: The paragraph presents the study-specific characterization and structural limitations without a local retraction warning.

The sentence the footnote ends: Measurements under extreme pressure are difficult, and in particular the elements are too light for an X-ray determination of crystal structure (X-ray crystallography).

Where the model says the article warns about this paper:

• In the opening paragraphs: in a Nature paper that was later retracted.
• In the opening paragraphs: The paper faced criticism due to its non-standard data analysis calling into question its conclusions, and in September 2022 it was retracted by Nature.
• Later (Claims of superconductivity): Two years later, the paper was retracted.
• Later (Criticism and retraction): Criticism and retraction
• Later (Criticism and retraction): On February 15, 2022, Nature added a cautionary Editor's Note to the article, and on 26 September 26, 2022, retracted the article entirely.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.88. Finding RW-39363 / 65584884 / #2, read from revision 1358563266 of Carbonaceous sulfur hydride.

The model's reasoning: These material-specific characterization claims derive from the study; independent sourcing is needed for their continued use.

The sentence the footnote ends: As of October 2020, the material's molecular structure remains uncharacterized, as extreme pressures and the light elements used are unsuitable for most measurements, such as X-ray determination.

Where the model says the article warns about this paper:

• In the opening paragraphs: in a Nature paper that was later retracted.
• In the opening paragraphs: The paper faced criticism due to its non-standard data analysis calling into question its conclusions, and in September 2022 it was retracted by Nature.
• Later (Claims of superconductivity): Two years later, the paper was retracted.
• Later (Criticism and retraction): Criticism and retraction
• Later (Criticism and retraction): On February 15, 2022, Nature added a cautionary Editor's Note to the article, and on 26 September 26, 2022, retracted the article entirely.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.90. Finding RW-39363 / 65584884 / #3, read from revision 1358563266 of Carbonaceous sulfur hydride.

The model's reasoning: This reports the paper's experimental synthesis claim without identifying the study as retracted in this paragraph.

Where the model says the article warns about this paper:

• In the opening paragraphs: in a Nature paper that was later retracted.
• In the opening paragraphs: The paper faced criticism due to its non-standard data analysis calling into question its conclusions, and in September 2022 it was retracted by Nature.
• Later (Claims of superconductivity): Two years later, the paper was retracted.
• Later (Criticism and retraction): Criticism and retraction
• Later (Criticism and retraction): On February 15, 2022, Nature added a cautionary Editor's Note to the article, and on 26 September 26, 2022, retracted the article entirely.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.97. Finding RW-39363 / 65584884 / #4, read from revision 1358563266 of Carbonaceous sulfur hydride.

The model's reasoning: The preparation details and stability observations are study-specific claims; source review is needed to assess whether the data concerns undermine them.

The sentence the footnote ends: It was reported that the crystal is not stable under 10 GPa (1.5 million psi) and can be destroyed if left at room temperature overnight.

Where the model says the article warns about this paper:

• In the opening paragraphs: in a Nature paper that was later retracted.
• In the opening paragraphs: The paper faced criticism due to its non-standard data analysis calling into question its conclusions, and in September 2022 it was retracted by Nature.
• Later (Claims of superconductivity): Two years later, the paper was retracted.
• Later (Criticism and retraction): Criticism and retraction
• Later (Criticism and retraction): On February 15, 2022, Nature added a cautionary Editor's Note to the article, and on 26 September 26, 2022, retracted the article entirely.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-39363 / 65584884 / #5, read from revision 1358563266 of Carbonaceous sulfur hydride.

The model's reasoning: The infobox repeats the study-associated proposed composition stated in the cited characterization paragraphs; an editor should check whether the infobox has independent support.

Where the model says the article warns about this paper:

• In the opening paragraphs: in a Nature paper that was later retracted.
• In the opening paragraphs: The paper faced criticism due to its non-standard data analysis calling into question its conclusions, and in September 2022 it was retracted by Nature.
• Later (Claims of superconductivity): Two years later, the paper was retracted.
• Later (Criticism and retraction): Criticism and retraction
• Later (Criticism and retraction): On February 15, 2022, Nature added a cautionary Editor's Note to the article, and on 26 September 26, 2022, retracted the article entirely.

The model's confidence: 0.85. Finding RW-39363 / 65584884 / #9, read from revision 1358563266 of Carbonaceous sulfur hydride.

The model's reasoning: This calculated property rests on the study-associated proposed formula; an editor should check whether an independent source establishes that composition.

Where the model says the article warns about this paper:

• In the opening paragraphs: in a Nature paper that was later retracted.
• In the opening paragraphs: The paper faced criticism due to its non-standard data analysis calling into question its conclusions, and in September 2022 it was retracted by Nature.
• Later (Claims of superconductivity): Two years later, the paper was retracted.
• Later (Criticism and retraction): Criticism and retraction
• Later (Criticism and retraction): On February 15, 2022, Nature added a cautionary Editor's Note to the article, and on 26 September 26, 2022, retracted the article entirely.

The model's confidence: 0.82. Finding RW-39363 / 65584884 / #10, read from revision 1358563266 of Carbonaceous sulfur hydride.

The model's reasoning: This sentence presents the retracted paper's findings about Bcr-Abl repression of PHLPP and reduced inhibitor efficacy after PHLPP depletion as established results, with no other supporting citation.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-44343 / 10292285 / #1, read from revision 1376190558 of PHLPP.

The model's reasoning: The sentence cites only this paper, but its title establishes the reconstruction topic and dates without revealing whether it documents student enrolment or industrial training; no abstract is available.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.90. Finding RW-22427 / 11535416 / #1, read from revision 1378186352 of North Koreans in Russia.

The model's reasoning: The UNC5-induced switch from attraction to repulsion and the UNC5/DCC receptor-complex mechanism state the retracted paper's findings as fact; the retraction identifies unreliable results and conclusions.

The sentence the footnote ends: The switch between attractive and repulsive netrin signaling can be mediated by misexpression of unc-5 in commissural axons.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-49334 / 34806886 / #1, read from revision 1355585651 of Tropic cues involved in growth cone guidance.

The model's reasoning: Both sentences present the paper's reported CaMKII–MAPK interaction and downstream p27 degradation as established findings; the retraction raises concerns about images and unavailable original data.

The sentence the footnote ends: Here, CaMKII activates MEK/ERK, which degrades the cell cycle arresting p27 protein

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-29239 / 55841611 / #1, read from revision 1377626591 of Calcium signaling in cell division.

The model's reasoning: This states the paper's finding that M. perstans prevalence increased significantly with age in the sampled Gabonese children.

The reference is tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-70140 / 3474864 / #1, read from revision 1349564938 of Acanthocheilonemiasis.

The model's reasoning: The adjacent disease-effect claim appears to generalize the paper's comparison of M. perstans prevalence in febrile and afebrile children; that comparison does not establish the broader claims about these named diseases.

The sentence the footnote ends: Recent studies focused on Gabon specifically, where febrile and tropical diseases are common.

Other citations on it:

• Simonsen, Paul E.; Onapa, Ambrose W.; Asio, Santa Maria (2011-09-01). "Mansonella perstans filariasis in Africa". Acta Tropica. The Diagnostics and Control of Neglected Tropical Helminth Diseases. 120, Supplement 1: S109–S120. doi:10.1016/j.actatropica.2010.01.014. PMID 20152790.

The reference is tagged {{Retracted}}.

The model's confidence: 0.85. Finding RW-70140 / 3474864 / #3, read from revision 1349564938 of Acanthocheilonemiasis.

The model's reasoning: The paragraph reports the paper's primitive-lineage, geographic-origin and anatomical findings without locally identifying them as discredited; the Guardian report plausibly supports the attributed origin claim but is reporting the same discovery.

The sentence the footnote ends: Further, the study concluded that the cheetahs originated in the Old World, not the New World as previously believed.

Where the model says the article warns about this paper:

• In the opening paragraphs: "Acinonyx kurteni", or the Linxia cheetah, is a discredited fossil specimen of an extinct cheetah discovered in China.
• In the opening paragraphs: After a long suspicion of the authenticity of the fossil, it was finally accepted as a forgery in 2012.
• Later (Discrediting): Tao Deng, Professor of Paleontology at the Institute for Vertebrate Paleontology and Paleoanthropology in China, upon proofreading the manuscript, noted the possible forgery of the fossil, and immediately reported this to PNAS.
• Later (Discrediting): Qiu Zhanxiang, a global expert in mammalian craniums, also supported the claim that the skull was forged.
• Later (Discrediting): In 2012, Deng was finally allowed access to the original fossil upon invoking the PNAS data access policy, and on examination, his primary observations were proven correct.
• Later (Discrediting): Deng asked Mazák to retract the paper. Mazák finally conceded by stating (on 20 August 2012 in PNAS) that: “after further examination, it was determined that the fossil used in the study was a composite specimen from the late Miocene laterite and not from the early Pleistocene loess. The article is hereby retracted.”

Other citations on it:

• Randerson, James (December 29, 2008). "Fossilised skull suggests cheetahs evolved in Asia not Americas". The Guardian. London. Retrieved April 23, 2010.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-5686 / 20938232 / #1, read from revision 1268142751 of Acinonyx kurteni.

The model's reasoning: The description presents anatomy, relative age and evolutionary implications derived from the composite specimen as valid, with no local retraction warning or other supporting citations.

The sentence the footnote ends: It is older than earlier described species, such as the European Acinonyx pardinensis (dated to about 2.2 Ma old) and the North African Acinonyx aicha (about 2.5 Ma old).

Where the model says the article warns about this paper:

• In the opening paragraphs: "Acinonyx kurteni", or the Linxia cheetah, is a discredited fossil specimen of an extinct cheetah discovered in China.
• In the opening paragraphs: After a long suspicion of the authenticity of the fossil, it was finally accepted as a forgery in 2012.
• Later (Discrediting): Tao Deng, Professor of Paleontology at the Institute for Vertebrate Paleontology and Paleoanthropology in China, upon proofreading the manuscript, noted the possible forgery of the fossil, and immediately reported this to PNAS.
• Later (Discrediting): Qiu Zhanxiang, a global expert in mammalian craniums, also supported the claim that the skull was forged.
• Later (Discrediting): In 2012, Deng was finally allowed access to the original fossil upon invoking the PNAS data access policy, and on examination, his primary observations were proven correct.
• Later (Discrediting): Deng asked Mazák to retract the paper. Mazák finally conceded by stating (on 20 August 2012 in PNAS) that: “after further examination, it was determined that the fossil used in the study was a composite specimen from the late Miocene laterite and not from the early Pleistocene loess. The article is hereby retracted.”

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-5686 / 20938232 / #2, read from revision 1268142751 of Acinonyx kurteni.

The model's reasoning: This repeats the hypothesis attributed to the retracted paper in the following paragraph without a local retraction warning; an editor should check whether this paragraph has an independent source, such as Mead's earlier books.

Where the model found it: Mead believes that welfare should be means-tested, so as to encourage behaviors among aid recipients to alleviate their poverty. According to Mead, extreme poverty is more commonly found in minority groups not because of a lack of work ethic. Rather, he argues that their belief in work unfortunately does not sufficiently bind their actions. This assessment is identical to JD Vance's assessment of poor, white Appalachians in his book Hillbilly Elegy. Hence, Mead argues "[t]he ideal in welfare... is to link benefits as tightly as possible to work. That requires a clear work test that employable recipients must meet as soon as they apply for aid, not sometime later. Equally important, there must be ample benefits to support working, particularly child and health care."

Where the model says the article warns about this paper:

• In the same section (Theories on welfare and poverty): NYU issued a statement expressing its rejection of "false, prejudicial, and stigmatizing assertions about the culture of communities of color in the United States" in his paper.
• In the same section (Theories on welfare and poverty): The journal publisher launched an investigation, and on July 31 the editor-in-chief retracted the paper and apologized for publishing it.

The model's confidence: 0.87. Finding RW-23869 / 22791353 / #3, read from revision 1360109387 of Lawrence Mead.

The model's reasoning: The paragraph reports the paper's numerical finding without explicitly identifying its retraction; methodological criticism does not clearly refute this specific result.

Where the model says the article warns about this paper:

• In the opening paragraphs: A report by WPI scientists of an association between CFS and XMRV was forcibly retracted by the journal Science when the results could not be replicated, and it was discovered that XMRV was a laboratory-created recombinant of two mouse viruses.
• In the same section (Science publication and response): Immediately after the Science paper publication, Patrick Moore and Masahiro Shuda wrote a highly critical F1000 dissent pointing out basic errors in the paper's methods, results and conclusions, "...this manuscript has flaws that leave the reader unsure of knowing precisely what was measured".
• In the same section (Science publication and response): In September 2011, the original authors published a "Partial Retraction" of their 2009 findings, in which they acknowledged that "some of the CFS peripheral blood mononuclear cell (PBMC) DNA preparations are contaminated with XMRV plasmid DNA." In December 2011, Science retracted the 2009 paper.
• In the same section (Contamination): Virologist John Coffin, who had previously expressed optimism about the initial WPI data, declared, "It's all contamination."
• In the same section (Controversial statements and criticism): The editors at Science in May 2011 requested that the authors of the 2009 Lombardi et al. study retract the article due to many independent studies that did not find an association of XMRV with CFS, and studies that have presented data showing contaminated laboratory reagents may have led to false positive results.
• In the same section (Controversial statements and criticism): The editor-in-chief of the journal Science subsequently published an expression of concern, "Because the validity of the study by Lombardi et al. is now seriously in question". In December 2011, Science retracted the 2009 paper.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.94. Finding RW-492 / 25732046 / #3, read from revision 1365452137 of Whittemore Peterson Institute.

The model's reasoning: This attributed CFS claim appears to extend the paper's proposed pathogenic role; an editor should check whether the paragraph's own sources adequately contextualize it.

Where the model found it: The Whittemore Peterson Institute and its employees have issued statements that have generated additional controversy. Some of these statements have been made to the press, while others are made in presentations to patient groups. Although there is no consensus that XMRV is a human pathogen or even associated with any disease, Mikovits has stated her opinion that the virus "undoubtedly causes some of the symptoms that are associated with" CFS. Elsewhere, however, she notes that causation has not been proven. According to Mikovits, XMRV "is clearly circulating through the population as is our fear and your fear" and has entered the US blood supply. She has also associated XMRV not only with CFS, but also with autism, Alzheimer's disease, multiple sclerosis, and other diseases. There is no evidence that XMRV is associated with these diseases: all studies that have looked for a connection have produced negative results. Virologist Vincent Racaniello has accused Mikovits of "inciting fear".

Where the model says the article warns about this paper:

• In the opening paragraphs: A report by WPI scientists of an association between CFS and XMRV was forcibly retracted by the journal Science when the results could not be replicated, and it was discovered that XMRV was a laboratory-created recombinant of two mouse viruses.
• In the same section (Science publication and response): Immediately after the Science paper publication, Patrick Moore and Masahiro Shuda wrote a highly critical F1000 dissent pointing out basic errors in the paper's methods, results and conclusions, "...this manuscript has flaws that leave the reader unsure of knowing precisely what was measured".
• In the same section (Science publication and response): In September 2011, the original authors published a "Partial Retraction" of their 2009 findings, in which they acknowledged that "some of the CFS peripheral blood mononuclear cell (PBMC) DNA preparations are contaminated with XMRV plasmid DNA." In December 2011, Science retracted the 2009 paper.
• In the same section (Contamination): Virologist John Coffin, who had previously expressed optimism about the initial WPI data, declared, "It's all contamination."
• In the same section (Controversial statements and criticism): The editors at Science in May 2011 requested that the authors of the 2009 Lombardi et al. study retract the article due to many independent studies that did not find an association of XMRV with CFS, and studies that have presented data showing contaminated laboratory reagents may have led to false positive results.
• In the same section (Controversial statements and criticism): The editor-in-chief of the journal Science subsequently published an expression of concern, "Because the validity of the study by Lombardi et al. is now seriously in question". In December 2011, Science retracted the 2009 paper.

The model's confidence: 0.79. Finding RW-492 / 25732046 / #4, read from revision 1365452137 of Whittemore Peterson Institute.

The model's reasoning: This states the retracted paper's antibody-response result as fact; the incomplete retraction reason requires checking the notice before deciding whether to remove it.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-5206 / 68892222 / #1, read from revision 1353163515 of Bacteriophage AP205.

The model's reasoning: This presents the paper's conclusion of superior immunogenicity as fact; the incomplete retraction reason requires checking the notice before deciding whether to remove it.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-5206 / 68892222 / #4, read from revision 1353163515 of Bacteriophage AP205.

The model's reasoning: This states the paper's virus-neutralization result as fact; the incomplete retraction reason requires checking the notice before deciding whether to remove it.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-5206 / 68892222 / #5, read from revision 1353163515 of Bacteriophage AP205.

The model's reasoning: The supplied abstract does not establish this Greek-origin finding; the other cited papers plausibly support it, so the retracted paper's contribution needs source inspection.

Where the model says the article warns about this paper:

• In the same section (Jews and Palestinians): Following a large number of complaints, the article was withdrawn from the journal and deleted from the scientific archive.
• In the same section (Jews and Palestinians): Academics who had already received a copy of the journal were urged to "physically remove" the article pages in a move that had no precedent in research publishing.
• In the same section (Jews and Palestinians): Andrew Goffey, a senior lecturer at Middlesex University, England, observing that "it was conceded that the article had not been removed on the basis of its scientific evidence," failed to find anything offensive in the paper.

Other citations on it:

• Arnaiz-Villena A, Gomez-Casado E, Martinez-Laso J (August 2002). "Population genetic relationships between Mediterranean populations determined by HLA allele distribution and a historic perspective". Tissue Antigens. 60 (2): 111–21. doi:10.1034/j.1399-0039.2002.600201.x. PMID 12392505.
• Arnaiz-Villena A, Dimitroski K, Pacho A, et al. (February 2001). "HLA genes in Macedonians and the sub-Saharan origin of the Greeks". Tissue Antigens. 57 (2): 118–27. doi:10.1034/j.1399-0039.2001.057002118.x. PMID 11260506. S2CID 652188.
• Arnaiz-Villena A, Martinez-Laso J, Alonso-Garciá J (September 2001). "The correlation between languages and genes: the Usko-Mediterranean peoples". Hum. Immunol. 62 (9): 1051–61. doi:10.1016/S0198-8859(01)00300-7. PMID 11543906.
• Arnaiz-Villena A, Karin M, Bendikuze N, et al. (April 2001). "HLA alleles and haplotypes in the Turkish population: relatedness to Kurds, Armenians and other Mediterraneans". Tissue Antigens. 57 (4): 308–17. doi:10.1034/j.1399-0039.2001.057004308.x. PMID 11380939.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.87. Finding RW-3030 / 10841641 / #3, read from revision 1375851211 of Antonio Arnaiz-Villena.

The model's reasoning: This migration explanation is absent from the supplied abstract, and the other cited papers may be its source.

Where the model says the article warns about this paper:

• In the same section (Jews and Palestinians): Following a large number of complaints, the article was withdrawn from the journal and deleted from the scientific archive.
• In the same section (Jews and Palestinians): Academics who had already received a copy of the journal were urged to "physically remove" the article pages in a move that had no precedent in research publishing.
• In the same section (Jews and Palestinians): Andrew Goffey, a senior lecturer at Middlesex University, England, observing that "it was conceded that the article had not been removed on the basis of its scientific evidence," failed to find anything offensive in the paper.

Other citations on it:

• Arnaiz-Villena A, Gomez-Casado E, Martinez-Laso J (August 2002). "Population genetic relationships between Mediterranean populations determined by HLA allele distribution and a historic perspective". Tissue Antigens. 60 (2): 111–21. doi:10.1034/j.1399-0039.2002.600201.x. PMID 12392505.
• Arnaiz-Villena A, Dimitroski K, Pacho A, et al. (February 2001). "HLA genes in Macedonians and the sub-Saharan origin of the Greeks". Tissue Antigens. 57 (2): 118–27. doi:10.1034/j.1399-0039.2001.057002118.x. PMID 11260506. S2CID 652188.
• Arnaiz-Villena A, Martinez-Laso J, Alonso-Garciá J (September 2001). "The correlation between languages and genes: the Usko-Mediterranean peoples". Hum. Immunol. 62 (9): 1051–61. doi:10.1016/S0198-8859(01)00300-7. PMID 11543906.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.87. Finding RW-3030 / 10841641 / #4, read from revision 1375851211 of Antonio Arnaiz-Villena.

The model's reasoning: The supplied abstract does not mention this Macedonian finding, while the other citation directly concerns Macedonian HLA relationships.

Where the model says the article warns about this paper:

• In the same section (Jews and Palestinians): Following a large number of complaints, the article was withdrawn from the journal and deleted from the scientific archive.
• In the same section (Jews and Palestinians): Academics who had already received a copy of the journal were urged to "physically remove" the article pages in a move that had no precedent in research publishing.
• In the same section (Jews and Palestinians): Andrew Goffey, a senior lecturer at Middlesex University, England, observing that "it was conceded that the article had not been removed on the basis of its scientific evidence," failed to find anything offensive in the paper.

Other citations on it:

• Arnaiz-Villena A, Dimitroski K, Pacho A, et al. (February 2001). "HLA genes in Macedonians and the sub-Saharan origin of the Greeks". Tissue Antigens. 57 (2): 118–27. doi:10.1034/j.1399-0039.2001.057002118.x. PMID 11260506. S2CID 652188.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.88. Finding RW-3030 / 10841641 / #6, read from revision 1375851211 of Antonio Arnaiz-Villena.

The model's reasoning: This presents the paper's conclusion that perforator flaps provide ideal tissue for breast reconstruction as fact; duplication does not itself invalidate that conclusion.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.98. Finding RW-2516 / 37494943 / #1, read from revision 1370411802 of Perforator flaps.

The model's reasoning: This relies on the paper's account of numerous suitable donor sites, extending it to near-universal eligibility despite the abstract's contraindications; a non-retracted source must support that scope.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.96. Finding RW-2516 / 37494943 / #2, read from revision 1370411802 of Perforator flaps.

The model's reasoning: These benefits elaborate the paper's muscle-preservation and reduced donor-site morbidity claims; the other reconstructive-surgery reviews plausibly support them, and the retraction concerns duplication.

Other citations on it:

• Granzow JW, Levine JL, Chiu ES, Allen RJ (November 2006). "Breast reconstruction using perforator flaps". J Surg Oncol. 94 (6): 441–54. doi:10.1002/jso.20481. PMID 17061279. S2CID 24974920.
• Neligan PC, Lipa JE (May 2006). "Perforator Flaps in Head and Neck Reconstruction". Semin Plast Surg. 20 (2): 56–63. doi:10.1055/s-2006-941711. PMC 2884771.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.94. Finding RW-2516 / 37494943 / #3, read from revision 1370411802 of Perforator flaps.

The model's reasoning: These asserted recovery benefits elaborate the paper's reduced donor-site morbidity conclusion; the cited review of perforator-flap advantages plausibly supports them, while the angiography source more directly concerns the following perfusion claim.

The sentence the footnote ends: Furthermore, patients have shown decreased postoperative pain and accelerated rehabilitation Nevertheless, there will always be a chance that the displaced tissue partially or completely dies considering the fact that the perfusion of the flap is difficult to assess intraoperatively.

Other citations on it:

• Guerra AB, Lyons GD, Dupin CL, Metzinger SE (July 2005). "Advantages of perforator flaps in reconstruction of complex defects of the head and neck". Ear Nose Throat J. 84 (7): 441–7. doi:10.1177/014556130508400717. PMID 16813036.
• Liu DZ, Mathes DW, Zenn MR, Neligan PC (July 2011). "The application of indocyanine green fluorescence angiography in plastic surgery". J Reconstr Microsurg. 27 (6): 355–64. doi:10.1055/s-0031-1281515. PMID 21717392. S2CID 260193952.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.88. Finding RW-2516 / 37494943 / #4, read from revision 1370411802 of Perforator flaps.

The model's reasoning: The reliable transfer and reduced donor-site morbidity benefits repeat the retracted paper's conclusion and the benefits cited later; an editor must check whether this paragraph has independent sourcing, particularly for sensory innervation.

Where the model found it: However, the health of the patient and possible contra-indications play an important role as well. Due to the development and improvement of cutaneous, myocutanous and fasciocutaneous tissue transpositions plastic surgeons are able to successfully restore the defect to its original shape. Nevertheless, functional recovery is not guaranteed in all patients. For the optimal renewal of shape and function, a suitable flap can be chosen to reconstruct the defect. In the case of using a so-called perforator flap, a reliable vascularization and the possibility of sensory (re) innervation can be combined with less donor-site morbidity and limited loss of function in the donor area.

The model's confidence: 0.85. Finding RW-2516 / 37494943 / #5, read from revision 1370411802 of Perforator flaps.

The model's reasoning: The citation attaches to this claim, but the title alone does not establish whether the paper argued for these specific dependencies and asymmetries; no abstract is available.

Other citations on it:

• Blackwell, Alan (2021). "Ethnographic artificial intelligence". Interdisciplinary Science Reviews. 46 (1–2): 198–211 – via Sage.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.88. Finding RW-73562 / 81801325 / #1, read from revision 1374628080 of Data colonialism.

The model's reasoning: This reports the therapy claim with an explicit retraction warning in the same paragraph, rather than presenting the cited study as reliable evidence.

The sentence the footnote ends: While most studies attempt to show the correlation between personality types and coronary heart disease, studies (that also later were questioned for non plausible results were considered unsafe, and were retracted in 2021) suggested that mental attitudes constitute an important prognostic factor for cancer and that as a method of treatment for cancer-prone patients, behavior therapy should be used.

Where the model says the article warns about this paper:

• In the same section (Other studies): studies (that also later were questioned for non plausible results were considered unsafe, and were retracted in 2021)

Other citations on it:

• Pelosi, Anthony J. (2019). "Personality and fatal diseases: Revisiting a scientific scandal". Journal of Health Psychology. 24 (4): 421–439. doi:10.1177/1359105318822045. ISSN 1359-1053. PMC 6712909. PMID 30791726.
• "King's College London enquiry into publications authored by Professor Hans Eysenck with Professor Ronald Grossarth-Maticek" (PDF). October 2019.
• Nigel Hawkes (2019), Works by eminent psychologist who doubted smoking caused cancer are "unsafe," finds inquiry

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.98. Finding RW-29959 / 1699268 / #1, read from revision 1376804284 of Type A and Type B personality theory.

The model's reasoning: Under the list introduction, this item asserts an interaction between SMARCC1 and BAZ1B (WSTF), relying solely on the retracted WINAC study.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-216 / 14760447 / #1, read from revision 1301028146 of SMARCC1.

The model's reasoning: The paragraph states the paper’s guide-binding, DNA-cleavage and PAM-independence findings as facts; the qualification about replication does not qualify all those mechanistic claims.

The sentence the footnote ends: Like the CRISPR/Cas system, NgAgo was reported by Chunyu Han et al. to be suitable for genome editing, but this has not been replicated.

Where the model says the article warns about this paper:

• Later (Controversy): The authors retracted the study in a statement published in Nature Biotechnology on 3 August 2017, citing the continued inability of the research community to replicate their results.
• Later (Controversy): In 2018, an investigation led by Han's university concluded that while Han's findings were flawed, he and his team did not intend to deceive the scientific community.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.98. Finding RW-8561 / 50686685 / #1, read from revision 1334892185 of NgAgo.

The model's reasoning: The paragraph presents the study’s claimed editing advantages and guide requirements without a local retraction warning; the terminal citation plausibly covers these preceding claims as well as the dosage comparison.

The sentence the footnote ends: The required dosage of ssDNA is less than that of a sgRNA expression plasmid.

Where the model says the article warns about this paper:

• Later (Controversy): The authors retracted the study in a statement published in Nature Biotechnology on 3 August 2017, citing the continued inability of the research community to replicate their results.
• Later (Controversy): In 2018, an investigation led by Han's university concluded that while Han's findings were flawed, he and his team did not intend to deceive the scientific community.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.97. Finding RW-8561 / 50686685 / #2, read from revision 1334892185 of NgAgo.

The model's reasoning: These claims repeat the paper's findings about EPCA-2's association with prostate cancer and superior diagnostic performance; an editor should check whether the paragraph has independent sources.

Where the model says the article warns about this paper:

• Later (EPCA-2 versus PSA): This paper has since been retracted by the publisher.
• Later (EPCA-2 versus PSA): In September 2008 the industry sponsor of EPCA-2, Onconome sued Dr Robert Getzenberg, JHU, and the University of Pittsburgh, his previous institution, claiming that Getzenberg misrepresented and falsified data related to EPCA-2 after Onconome sponsored 13 million dollars of research over five years in Getzenberg's labs at University of Pittsburgh and Johns Hopkins for a blood test for prostate cancer.

The model's confidence: 0.98. Finding RW-1844 / 6928455 / #2, read from revision 1367468495 of Early prostate cancer antigen-2.

The model's reasoning: The paragraph presents the paper's procedure and claimed successful, rejection-free transplant as established facts, extending its four-month observation to an ongoing absence of rejection.

The sentence the footnote ends: The method involves obtaining a donor organ, removing the cells and MHC antigens from the donor organ, and colonizing it with stem cells obtained from the patient.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-48567 / 25671436 / #1, read from revision 1374409703 of Neo-organ.

The model's reasoning: The paragraph repeats the paper's claimed successful transplant and generalizes its claimed freedom from rejection; fabrication undermines these claims.

The sentence the footnote ends: This process was used in the first successful stem-cell based organ transplant by removing the cells and MHC antigens from the donor organ.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-48567 / 25671436 / #2, read from revision 1374409703 of Neo-organ.

The model's reasoning: The opening summarizes the clinical success and rejection-reduction claims illustrated below by the retracted airway study; an editor should check whether this paragraph has independent sources supporting that generalization.

Where the model found it: In tissue engineering, neo-organ is the final structure of a procedure based on transplantation consisting of endogenous stem/progenitor cells grown ex vivo within predesigned matrix scaffolds. Current organ donation faces the problems of patients waiting to match for an organ and the possible risk of the patient's body rejecting the organ. Neo-organs are being researched as a solution to those problems with organ donation. Suitable methods for creating neo-organs are still under development. One experimental method is using adult stem cells, which use the patients own stem cells for organ donation. Currently this method can be combined with decellularization, which uses a donor organ for structural support but removes the donors cells from the organ. Similarly, the concept of 3-D bioprinting organs has shown experimental success in printing bioink layers that mimic the layer of organ tissues. However, these bioinks do not provide structural support like a donor organ. Current methods of clinically successful neo-organs use a combination of decellularized donor organs, along with adult stem cells of the organ recipient to account for both the structural support of a donor organ and the personalization of the organ for each individual patient to reduce the chance of rejection.

The model's confidence: 0.90. Finding RW-48567 / 25671436 / #3, read from revision 1374409703 of Neo-organ.

The model's reasoning: The claim of clinical success and reduced rejection generalizes the organ-replacement outcome discussed in the retracted paper.

The reference is tagged {{Retracted}}.

The model's confidence: 0.91. Finding RW-72108 / 25671436 / #2, read from revision 1374409703 of Neo-organ.

The model's reasoning: This presents the first stem-cell-based organ replacement discussed by the paper as a successful clinical outcome.

The reference is tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-72108 / 25671436 / #3, read from revision 1374409703 of Neo-organ.

The model's reasoning: This explicitly asserts successful surgery and absence of rejection for the organ replacement discussed in the retracted paper.

The reference is tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-72108 / 25671436 / #5, read from revision 1374409703 of Neo-organ.

The model's reasoning: This repeats the same first-transplant success claim cited to the retracted paper in the Adult Stem Cells section; an editor must check whether this paragraph has an independent source.

Where the model found it: Researchers have begun to focus on decellularization for organ transplants since it reduces the chance of rejection to almost none. This process was used in the first successful stem-cell based organ transplant by removing the cells and MHC antigens from the donor organ. There are different ways to remove the cells from the organ which can include physical, chemical, and enzymatic treatments. This method is especially useful when trying to create a neo-heart because the heart needs to be created in a way where the structure remains. Since the stem cells used are currently not able to maintain a shape, researchers have started to look more into decellularization of existing organs to be able to perform successful transplant procedures without the problem of rejection. While this method may assist with the problem of rejection, donors are still needed to provide this structure to patients.

The model's confidence: 0.95. Finding RW-72108 / 25671436 / #6, read from revision 1374409703 of Neo-organ.

The model's reasoning: The item reports the paper's retraction, not its findings on MSC attachment and proliferation on PET and PU scaffolds.

Where the model says the article warns about this paper:

• In the same section (Retractions): The following papers authored by Macchiarini have been retracted:
• In the same section (Retractions): January 2024: a 2012 Biomaterials paper retracted for "falsification and/or fabrication".

Other citations on it:

• Gustafsson, Ylva; Haag, Johannes; Jungebluth, Philipp; Lundin, Vanessa; Ling Lim, Mei; Baiguera, Silvia; Ajalloueian, Fatemeh; Del Gaudio, Costantino; Bianco, Alessandra; Moll, Guido; Sjöqvist, Sebastian; Lemon, Greg; Teixeira, Ana Isabel; Macchiarini, Paolo (4 January 2024). "Retraction notice to "Viability and proliferation of rat MSCs on adhesion protein-modified PET and PU scaffolds" [Biomaterials 33/12 (2012) 8094–8103]". Biomaterials. 305 122445. doi:10.1016/j.biomaterials.2023.122445. PMID 38176982.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-50888 / 24365639 / #1, read from revision 1375891403 of Paolo Macchiarini.

The model's reasoning: This describes the original transplant procedure, not the retracted follow-up findings about graft function, safety or patient recovery; a non-retracted source should support this background.

Where the model says the article warns about this paper:

• Later (Retractions): The following papers authored by Macchiarini have been retracted:
• Later (Retractions): October 2023, 2013 Lancet paper, retracted by the journal for containing fabrication and falsification in several places, and three falsified figures:

Other citations on it:

• Cowell, Alan; Grady, Denise (19 November 2008). "Europeans Announce Pioneering Surgery". The New York Times.
• Macchiarini, Paolo; Jungebluth, Philipp; Go, Tetsuhiko; Asnaghi, M Adelaide; Rees, Louisa E; Cogan, Tristan A; Dodson, Amanda; Martorell, Jaume; Bellini, Silvia; Parnigotto, Pier Paolo; Dickinson, Sally C; Hollander, Anthony P; Mantero, Sara; Conconi, Maria Teresa; Birchall, Martin A (December 2008). "Retracted: Clinical transplantation of a tissue-engineered airway". The Lancet. 372 (9655): 2023–2030. doi:10.1016/S0140-6736(08)61598-6. PMID 19022496. S2CID 13153058. (Retracted, see doi:10.1016/S0140-6736(23)02341-3, PMID 37898523, Retraction Watch)

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.92. Finding RW-48568 / 24365639 / #1, read from revision 1375891403 of Paolo Macchiarini.

The model's reasoning: The paragraph reports the study's conclusion but immediately explains its retraction and the authors' admission that their analyses could not address racial disparities in shooting probability.

Where the model says the article warns about this paper:

• In the same section (Allegations of use of excessive force by police): The authors later retracted the paper because although "our data and statistical approach were appropriate for investigating whether officer characteristics are related to the race of civilians fatally shot by police," the paper had been "cited as providing support for the idea that there are no racial biases in fatal shootings, or policing in general" whereas in fact their analyses "are inadequate to address racial disparities in the probability of being shot."

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-23720 / 44751865 / #1, read from revision 1378216353 of Black Lives Matter.

The model's reasoning: These stellar-orbit and age estimates are not identified in the supplied abstract as findings of the retracted analysis; the later paper may independently establish them.

Where the model says the article warns about this paper:

• In the opening paragraphs: A Jupiter-mass planet or a low-mass brown dwarf was once thought to be orbiting the central star, but this has been ruled out.
• Later (Claims of a planetary system): Artist's impression of HD 131399 Ab, before it was found to be a background star.
• Later (Claims of a planetary system): After subsequent data published in 2022 confirmed that the object is a background star, the paper announcing the putative discovery was retracted.

Other citations on it:

• Nielsen, Eric L.; et al. (2017). "Evidence that the Directly-Imaged Planet HD 131399 Ab is a Background Star". The Astronomical Journal. 154 (6): 218. arXiv:1705.06851. Bibcode:2017AJ....154..218N. doi:10.3847/1538-3881/aa8a69. S2CID 55138870.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.75. Finding RW-36523 / 54629028 / #6, read from revision 1377637721 of HD 131399.

The model's reasoning: The attributed classification and mass repeat the retracted planetary characterization; the later dynamical paper plausibly supports the conditional fate, not an independent planetary identification.

Where the model says the article warns about this paper:

• In the opening paragraphs: A Jupiter-mass planet or a low-mass brown dwarf was once thought to be orbiting the central star, but this has been ruled out.
• In the same section (Claims of a planetary system): Artist's impression of HD 131399 Ab, before it was found to be a background star.
• In the same section (Claims of a planetary system): After subsequent data published in 2022 confirmed that the object is a background star, the paper announcing the putative discovery was retracted.

Other citations on it:

• Veras, Dimitri; Mustill, Alexander J.; Gänsicke, Boris T. (2017). "The unstable fate of the planet orbiting the a star in the HD 131399 triple stellar system". Monthly Notices of the Royal Astronomical Society. 465 (2): 1499. arXiv:1611.00007. Bibcode:2017MNRAS.465.1499V. doi:10.1093/mnras/stw2821. S2CID 73723946.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.96. Finding RW-36523 / 54629028 / #9, read from revision 1377637721 of HD 131399.

The model's reasoning: The list presents a CHAF1B–BAZ1B (WSTF) interaction supported only by the retracted WINAC study, whose data and conclusions were affected.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.96. Finding RW-216 / 14774943 / #1, read from revision 1328170928 of CHAF1B.

The model's reasoning: This reports the review's conclusion as part of its ghostwriting history, and the same paragraph explicitly identifies the paper as retracted.

Where the model says the article warns about this paper:

• In the same section (Ghostwriting): The paper, in Regulatory Toxicology and Pharmacology, was retracted on November 28, 2025 by the editor-in-chief, citing "lack of authorial independence", "misrepresentation of contributions", "questions of financial compensation", and a variety of concerns over whether the conclusions presented were adequately supported by the scientific data.

Other citations on it:

• Foucart, Stéphane (December 3, 2025). "Influential study on glyphosate safety retracted 25 years after publication". Le Monde. Retrieved December 4, 2025.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-68223 / 93825 / #1, read from revision 1377165553 of Monsanto.

The model's reasoning: The sentence describes the paper's retraction and concerns about its conclusions rather than relying on its safety finding.

Where the model says the article warns about this paper:

• In the same section (Ghostwriting): The paper, in Regulatory Toxicology and Pharmacology, was retracted on November 28, 2025 by the editor-in-chief, citing "lack of authorial independence", "misrepresentation of contributions", "questions of financial compensation", and a variety of concerns over whether the conclusions presented were adequately supported by the scientific data.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-68223 / 93825 / #2, read from revision 1377165553 of Monsanto.

The model's reasoning: This describes the paper's misconduct controversy and retraction, rather than presenting its finding about herbivory and developmental instability as true.

Where the model says the article warns about this paper:

• In the same section (Flora and fauna): Prior to his work at Chernobyl, Møller was accused of falsifying data in a 1998 paper about asymmetry in oak leaves, which he retracted in 2001.

Other citations on it:

• Vogel, Gretchen; Proffitt, Fiona; Stone, Richard (28 January 2004). "Ecologists Rocked by Misconduct Finding". Science. Retrieved 17 February 2022.
• Borrell, Brendan (2007). "A Fluctuating Reality: Accused of fraud, Anders Pape Møller has traveled from superstar evolutionary biologist to pariah" (PDF). The Scientist. 21 (1): 26–. Archived from the original (PDF) on 17 February 2022. Retrieved 17 February 2022.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-22709 / 1021628 / #1, read from revision 1370473101 of Chernobyl exclusion zone.

The model's reasoning: This explicitly retells the identified paper's conclusions without a local retraction notice; an editor should check whether the item's unseen footnotes independently support this account.

Where the model found it: In the WR, "we judge that there was too much reliance on peer review, which was not necessarily independent." It continued, a "social network of authorships in temperature reconstruction is described of at least 43 authors with direct ties to Mann by virtue of having coauthored papers with him. The findings from this analysis suggest that authors in the area of paleoclimate studies are closely connected and thus ‘independent studies’ may not be as independent as they might appear on the surface. At the hearings, Wegman stated this was a "hypothesis", and "should be taken with a grain of salt." He listed 6 people who had participated in his own informal peer review process via email after the report was finalized and said they had no objection to the subcommittee submitting it. He defined the social network as peer reviewers that had "actively collaborated with him in writing research papers" and answered that none of his peer reviewers had. One of those listed by Wegman was the statistician Grace Wahba, who told North by email "Hey they used my name and they said I was a referee. He sent it to me about 3 days beforehand and I sent him a bunch of criticisms which they didn‘t take into account." At the hearings, North criticised aspects of this analysis, which had not been available for the NRC to examine. He "was not impressed by the social network analysis" and differed from "the report's conclusions on this subject". Indeed, he would look favourably on an application for tenure by a young scientist who "got busy and found himself 43 coauthors." In John Quiggin's opinion, the social network analysis was not based on meaningful criteria, did not prove a conflict of interest and did not apply at the time of the 1998 and 1999 publications. Such a network of co-authorship is not unusual in narrowly defined areas of science. When allegations of plagiarism in this section were examined, Wegman said that material in it had been "basically copied and pasted" by a student who was the "most knowledgeable" person about such analyses on his team, as she had taken a one-week course on network analysis with Kathleen Carley of Carnegie Mellon University. Carley described the paper based on this section as "more of an opinion piece", lacking the data needed to support its argument. The paper based on this social network analysis, published by Wegman and his former student Said, reached the conclusion that eminent scientists ought not work together, and that the findings of studies would be less biased when a "principal author tends to co-author papers with younger colleagues who were his students".

Where the model says the article warns about this paper:

• In the same section (Social network analysis paper): Wegman and Said published a paper (since retracted) in 2008 in the journal Computational Statistics & Data Analysis
• In the same section (Social network analysis paper): On 16 March 2011, Wegman sent an email to the journal saying that a student "had basically copied and pasted" work by other authors into the Wegman Report, and this text had been used in the journal paper without acknowledgement.
• In the same section (Social network analysis paper): In May 2011 the journal's editor, Stanley Azen of the University of Southern California, announced that the journal was retracting the paper, because it used portions of other authors' writings without sufficient attribution.
• In the same section (Social network analysis paper): Network analysis expert Kathleen Carley described it as an opinion piece which speculated that collaboration between scientists "leads to peer review abuse. No data is provided to support this argument".
• In the same section (Disciplinary procedure): A Nature editorial commented on the implication that the plagiarised material in the retracted paper was likely to also be present in the earlier "infamous" Wegman Report, including allegations against Mann and his co-authors which had frequently been cited by climate-change deniers.
• In the same section (Disciplinary procedure): He said that the 2008 social network analysis paper was investigated by a separate committee which unanimously found "that plagiarism occurred in contextual sections of the (CSDA) article, as a result of poor judgment for which Professor Wegman, as team leader, must bear responsibility", and Wegman was to receive an "official letter of reprimand".

The model's confidence: 0.96. Finding RW-6207 / 38501018 / #3, read from revision 1360746292 of Wegman Report.

The model's reasoning: These mechanism claims closely match the retracted paper's abstract; the other citation plausibly supports lithium–neuropeptide synergy, and the retraction was for duplication rather than invalid findings.

The sentence the footnote ends: In addition, by inhibiting NMDA receptors, lithium ions can regulate calcium homeostasis and inhibit the activation of calcium-dependent apotosis, also showing synergy with neuropeptides.

Other citations on it:

• Gromova, O. A.; Torshin, I. Yu.; Gogoleva, I. V.; Pronin, A. V.; Stelmashuk, E. V.; Isaev, N. K.; Genrikhs, E. E.; Demidov, V. I.; Volkov, A. Yu.; Khaspekov, G. L.; Alexandrova, O. P. (2015). "Pharmacokinetic and pharmacodynamic synergism between neuropeptides and lithium in the neurotrophic and neuroprotective action of cerebrolysin". Zhurnal Nevrologii I Psikhiatrii Im. S.s. Korsakova. 115 (3): 65–72. doi:10.17116/jnevro20151153165-72. PMID 26120985.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-27844 / 83484354 / #1, read from revision 1370036627 of Lithium ascorbate.

The model's reasoning: The paper reports VDR interactions with WINAC, but its abstract does not establish a MED1–VDR interaction; the other cited study plausibly supports nuclear receptor interactions independently.

Other citations on it:

• Ito M, Yuan CX, Malik S, Gu W, Fondell JD, Yamamura S, Fu ZY, Zhang X, Qin J, Roeder RG (March 1999). "Identity between TRAP and SMCC complexes indicates novel pathways for the function of nuclear receptors and diverse mammalian activators". Mol. Cell. 3 (3): 361–70. doi:10.1016/s1097-2765(00)80463-3. PMID 10198638.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.90. Finding RW-216 / 14723679 / #1, read from revision 1300842504 of MED1.

The model's reasoning: The abstract does not address MED1–estrogen receptor interactions, so this citation's relevance is uncertain; the other cited paper explicitly reports the interaction through TRAP220.

Other citations on it:

• Kang YK, Guermah M, Yuan CX, Roeder RG (March 2002). "The TRAP/Mediator coactivator complex interacts directly with estrogen receptors alpha and beta through the TRAP220 subunit and directly enhances estrogen receptor function in vitro". Proc. Natl. Acad. Sci. U.S.A. 99 (5): 2642–7. Bibcode:2002PNAS...99.2642K. doi:10.1073/pnas.261715899. PMC 122401. PMID 11867769.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.90. Finding RW-216 / 14723679 / #2, read from revision 1300842504 of MED1.

The model's reasoning: This repeats the breast cancer overexpression claim described in the retracted paper's abstract as prior work; an editor must check whether this paragraph has an independent source for these specific observations.

Where the model found it: In breast cancer, HYAL1 is also overexpressed in cell lines MDA-MB-231 and MCF-7 and invasive duct cancer tissues and metastatic lymph nodes. Higher HYAL1 expression has also been detected in primary tumor tissue from patients with subsequent brain metastases versus those without.

The model's confidence: 0.78. Finding RW-40388 / 14798929 / #2, read from revision 1315047393 of HYAL1.

The model's reasoning: This states the retracted paper's finding of distinct AP- and EPSP-induced NPAS4 mechanisms as fact, without a retraction warning in this paragraph.

Where the model says the article warns about this paper:

• In the same section (Career and research): In 2024, a paper from Bloodgood's lab in Cell Journal was retracted after reports of manipulated data. The paper's first author Gian-Stefano Brigidi, admitted to falsifying images and fabricating experiments, and multiple investigations — by UCSD, the University of Utah, and the U.S. Office of Research Integrity — concluded that he had engaged in research misconduct.
• In the same section (Career and research): She co-signed the retraction notice and took the difficult steps of notifying NIH, the journal, and institutional leadership, while also leading her trainees through the scientific and personal fallout of the scientific misconduct.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-54183 / 63493496 / #1, read from revision 1372224137 of Brenda Bloodgood.

The model's reasoning: This presents the retracted paper's stimulus-specific NPAS4 heterodimer and gene-regulation findings as fact; no other source is cited on the sentence.

Where the model says the article warns about this paper:

• In the same section (Career and research): In 2024, a paper from Bloodgood's lab in Cell Journal was retracted after reports of manipulated data. The paper's first author Gian-Stefano Brigidi, admitted to falsifying images and fabricating experiments, and multiple investigations — by UCSD, the University of Utah, and the U.S. Office of Research Integrity — concluded that he had engaged in research misconduct.
• In the same section (Career and research): She co-signed the retraction notice and took the difficult steps of notifying NIH, the journal, and institutional leadership, while also leading her trainees through the scientific and personal fallout of the scientific misconduct.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-54183 / 63493496 / #2, read from revision 1372224137 of Brenda Bloodgood.

The model's reasoning: Although the footnote ends at the publication details, the paragraph also presents the paper's safety and immunogenicity conclusions without a local retraction warning.

The sentence the footnote ends: The first clinical trial led by Salimian and Amani, and supervised by Hassan Abolghasemi and Gholamhossein Alishiri, was published in the Journal of Medical Virology 27 August 2022.

Where the model says the article warns about this paper:

• In the opening paragraphs: The clinical trial reported in the Journal of Medical Virology met a critical scrutiny in May 2023, after which the journal declared it as an invalid study and retracted it in March 2024.
• Later (Controversy): Forthal pointed out several data inconsistency, misrepresentation, vague experimental methods and conflict of interest that would have used to reject the paper in the first place. He made a concern that "a manuscript containing so many serious flaws would have been accepted for publication following peer review, and given these issues, a retraction may be in order". As the criticism was posted on PubPeer, other scientists also raised other concerns in the study. Australian epidemiologist Gideon Meyerowitz-Katz stated that the report had several other "impossible" and "contradictory" results.
• Later (Controversy): With no further response from the authors, the journal announced retraction of the paper on 2 March 2024, with a note:
• Later (Controversy): The retraction has been agreed due to concerns raised by third parties regarding issues with the data presented in the article. Several inconsistencies concerning the information provided about the analyzed subjects were additionally identified. Furthermore, the authors failed to disclose the presence of potential conflicts of interest that may have affected the interpretation of the results presented. Accordingly, the editors consider the conclusions of this manuscript to be invalid. The authors have been informed of the decision to retract but did not agree with it.
• Later (Controversy): One of the supervisors of the clinical trial, Abolghasemi claimed that the criticism and retraction were not of scientific causes but of prejudices, and his team was not given a chance for explanation.

Other citations on it:

• "Clinical studies of Noora vaccine published in Medical Virology". Tehran Times. 12 October 2022. Retrieved 18 June 2024.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-52438 / 68602473 / #1, read from revision 1339032037 of Noora (vaccine).

The model's reasoning: The quotation repeats the retracted paper's safety, immune-response and dose conclusions without identifying the retraction in this passage.

The sentence the footnote ends: The dose of 80 μg is an appropriate dose for injection in the next phases of the trial.

Where the model says the article warns about this paper:

• In the opening paragraphs: The clinical trial reported in the Journal of Medical Virology met a critical scrutiny in May 2023, after which the journal declared it as an invalid study and retracted it in March 2024.
• Later (Controversy): Forthal pointed out several data inconsistency, misrepresentation, vague experimental methods and conflict of interest that would have used to reject the paper in the first place. He made a concern that "a manuscript containing so many serious flaws would have been accepted for publication following peer review, and given these issues, a retraction may be in order". As the criticism was posted on PubPeer, other scientists also raised other concerns in the study. Australian epidemiologist Gideon Meyerowitz-Katz stated that the report had several other "impossible" and "contradictory" results.
• Later (Controversy): With no further response from the authors, the journal announced retraction of the paper on 2 March 2024, with a note:
• Later (Controversy): The retraction has been agreed due to concerns raised by third parties regarding issues with the data presented in the article. Several inconsistencies concerning the information provided about the analyzed subjects were additionally identified. Furthermore, the authors failed to disclose the presence of potential conflicts of interest that may have affected the interpretation of the results presented. Accordingly, the editors consider the conclusions of this manuscript to be invalid. The authors have been informed of the decision to retract but did not agree with it.
• Later (Controversy): One of the supervisors of the clinical trial, Abolghasemi claimed that the criticism and retraction were not of scientific causes but of prejudices, and his team was not given a chance for explanation.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-52438 / 68602473 / #2, read from revision 1339032037 of Noora (vaccine).

The model's reasoning: This attributes an immunity finding to the same Phase I trial, although the supplied abstract does not establish Omicron-specific protection; an editor should check any independent source attached here.

Where the model found it: Noora vaccine requires three doses given by intramuscular injection on days 0, 21 and 35. Phase I clinical trial was claimed to indicate successful immunity against SARS-CoV-2 Omicron variant.

Where the model says the article warns about this paper:

• In the opening paragraphs: The clinical trial reported in the Journal of Medical Virology met a critical scrutiny in May 2023, after which the journal declared it as an invalid study and retracted it in March 2024.
• Later (Controversy): Forthal pointed out several data inconsistency, misrepresentation, vague experimental methods and conflict of interest that would have used to reject the paper in the first place. He made a concern that "a manuscript containing so many serious flaws would have been accepted for publication following peer review, and given these issues, a retraction may be in order". As the criticism was posted on PubPeer, other scientists also raised other concerns in the study. Australian epidemiologist Gideon Meyerowitz-Katz stated that the report had several other "impossible" and "contradictory" results.
• Later (Controversy): With no further response from the authors, the journal announced retraction of the paper on 2 March 2024, with a note:
• Later (Controversy): The retraction has been agreed due to concerns raised by third parties regarding issues with the data presented in the article. Several inconsistencies concerning the information provided about the analyzed subjects were additionally identified. Furthermore, the authors failed to disclose the presence of potential conflicts of interest that may have affected the interpretation of the results presented. Accordingly, the editors consider the conclusions of this manuscript to be invalid. The authors have been informed of the decision to retract but did not agree with it.
• Later (Controversy): One of the supervisors of the clinical trial, Abolghasemi claimed that the criticism and retraction were not of scientific causes but of prejudices, and his team was not given a chance for explanation.

The model's confidence: 0.90. Finding RW-52438 / 68602473 / #3, read from revision 1339032037 of Noora (vaccine).

The model's reasoning: These physical and atmospheric properties repeat the retracted planet characterization; past-tense attribution does not explain that it was refuted.

The sentence the footnote ends: Its atmosphere was shown to contain both water and methane through the use of near-infrared spectroscopy (1.4–1.6 μm).

Where the model says the article warns about this paper:

• In the opening paragraphs: HD 131399 Ab (nicknamed Scorpion-1b) is a K-type or M-type main-sequence star located in the constellation of Centaurus which was mistaken for an exoplanet orbiting the quadruple-star system HD 131399.
• Later (Discovery): In May 2017, observations collected with the Gemini Planet Imager, and including a reanalysis of the SPHERE data, suggested that this target is in fact a background star. Further observations in 2022 confirmed this.

Other citations on it:

• "Astronomers spot giant planet with 3 suns". CBC News. The Associated Press. 7 July 2016. Retrieved 10 May 2026.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-36523 / 51021135 / #1, read from revision 1373909809 of HD 131399 Ab.

The model's reasoning: The orbital period and predicted illumination depend on the erroneous interpretation of the object as a planet in the system.

The sentence the footnote ends: One orbit of HD 131399 Ab was thought to take 550 years.

Where the model says the article warns about this paper:

• In the opening paragraphs: HD 131399 Ab (nicknamed Scorpion-1b) is a K-type or M-type main-sequence star located in the constellation of Centaurus which was mistaken for an exoplanet orbiting the quadruple-star system HD 131399.
• Later (Discovery): In May 2017, observations collected with the Gemini Planet Imager, and including a reanalysis of the SPHERE data, suggested that this target is in fact a background star. Further observations in 2022 confirmed this.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.98. Finding RW-36523 / 51021135 / #2, read from revision 1373909809 of HD 131399 Ab.

The model's reasoning: The claimed 80 AU orbit repeats the retracted planet interpretation; an editor should check whether this paragraph has an independent source addressing the refutation.

Where the model found it: The central star HD 131399A is orbited by the less massive stars (HD 131399B and HD 131399C) at about 300 astronomical units (AU). HD 131399B and HD 131399C orbit each other at about 10 AU, while HD 131399 Ab was thought to orbit HD 131399A at about 80 AU.

Where the model says the article warns about this paper:

• In the opening paragraphs: HD 131399 Ab (nicknamed Scorpion-1b) is a K-type or M-type main-sequence star located in the constellation of Centaurus which was mistaken for an exoplanet orbiting the quadruple-star system HD 131399.
• Later (Discovery): In May 2017, observations collected with the Gemini Planet Imager, and including a reanalysis of the SPHERE data, suggested that this target is in fact a background star. Further observations in 2022 confirmed this.

The model's confidence: 0.97. Finding RW-36523 / 51021135 / #4, read from revision 1373909809 of HD 131399 Ab.

The model's reasoning: The caption presents the retracted planetary orbit without qualification; an editor should check the illustration's sourcing and context.

Where the model says the article warns about this paper:

• In the opening paragraphs: HD 131399 Ab (nicknamed Scorpion-1b) is a K-type or M-type main-sequence star located in the constellation of Centaurus which was mistaken for an exoplanet orbiting the quadruple-star system HD 131399.
• Earlier (Discovery): In May 2017, observations collected with the Gemini Planet Imager, and including a reanalysis of the SPHERE data, suggested that this target is in fact a background star. Further observations in 2022 confirmed this.

The model's confidence: 0.95. Finding RW-36523 / 51021135 / #5, read from revision 1373909809 of HD 131399 Ab.

The model's reasoning: The separately presented quotation explicitly attributes this claim to Emil’s letter; it repeats that letter’s central argument, not an independently sourced finding.

Where the model found it: “Don’t use an instrument of oppression as a symbol of diversity and inclusion […] It has become “liberal” to see the hijab as a symbol of equity, diversity and inclusion. […] But respect does not alter the fact that the hijab, the niqab and the burka are also instruments of oppression for millions of girls and women around the world who are not allowed to make a choice”.

Where the model says the article warns about this paper:

• In the opening paragraphs: This campaign has started when a letter written by Dr Sherif Emil—a Canadian Children’s surgeon—and published in the Canadian Medical Association Journal, in which he criticizes promotion of hijab as a symbol of diversity, was retracted due to the accusations of Islamophobia.
• In the same section (Creation): The letter got retracted and the editor of CMAJ apologized for publishing a “wrong, hurtful, and offensive” letter.

The model's confidence: 0.94. Finding RW-34844 / 69797747 / #2, read from revision 1333292424 of LetUsTalk.

The model's reasoning: The caption describes the paper's flawed illustration and explicitly identifies the paper as retracted; it does not endorse its scientific findings.

Where the model says the article warns about this paper:

• In the same section (Notable usage and controversy): An anatomically incorrect diagram of a rat's penis and testicles illustrated by Midjourney, published in a now-retracted Frontiers in Cell and Developmental Biology paper, which went viral in 2024
• In the same section (Notable usage and controversy): The paper was retracted a day after the images went viral on Twitter.

Other citations on it:

• Schultz, Isaac (February 15, 2024). "'Rat Dck' Among Gibberish AI Images Published in Science Journal". Gizmodo. Archived from the original on February 15, 2024. Retrieved February 16, 2024.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-51983 / 71151961 / #1, read from revision 1376498512 of Midjourney.

The model's reasoning: The paragraph reports the paper's publication controversy and explicitly states its retraction, rather than relying on its scientific findings.

Where the model says the article warns about this paper:

• In the same section (Notable usage and controversy): An anatomically incorrect diagram of a rat's penis and testicles illustrated by Midjourney, published in a now-retracted Frontiers in Cell and Developmental Biology paper, which went viral in 2024
• In the same section (Notable usage and controversy): The paper was retracted a day after the images went viral on Twitter.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-51983 / 71151961 / #2, read from revision 1376498512 of Midjourney.

The model's reasoning: These sentences report the review's findings on long-term benefit, comparison with glucocorticoids, adjunctive use and adverse events; the review was withdrawn as out of date.

The sentence the footnote ends: Reported adverse events are minor and temporary.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-32790 / 44940009 / #1, read from revision 1372807261 of Surgery for temporomandibular joint dysfunction.

The model's reasoning: These sentences report the withdrawn review's pain comparison and conclusions; the research-only recommendation is stronger than the supplied abstract, and withdrawal as out of date calls for a current source.

The sentence the footnote ends: The review concluded by suggesting that arthrocentesis only be used for TMD within well designed randomized controlled trial (i.e. for the purposes of further research and not for routine management).

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-31698 / 44940009 / #5, read from revision 1372807261 of Surgery for temporomandibular joint dysfunction.

The model's reasoning: This describes authorship, retraction, and methodological criticism, rather than presenting the review's findings as true.

Where the model says the article warns about this paper:

• In the same section (COVID-19 misinformation): He is a co-author of a 2023 review, "A systematic review of autopsy findings in deaths after COVID-19 vaccination", that was retracted and heavily criticized for its methodological errors.

Other citations on it:

• Carballo-Carbajal, Iria (July 31, 2023). "Flawed preprint based on autopsies inadequate to demonstrate that COVID-19 vaccines caused 74% of those deaths". Health Feedback. Retrieved March 15, 2026.
• McDonald, Jessica (July 5, 2024). "Flawed Autopsy 'Review' Revives Unsupported Claims of COVID-19 Vaccine Harm, Censorship". FactCheck.org. Retrieved June 2, 2025.
• Roley, Gwen (July 6, 2023). "Misinformation swirls around unpublished paper on Covid-19 vaccine risks". Fact Check. Retrieved June 2, 2025.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-55662 / 68179014 / #1, read from revision 1369762884 of Peter A. McCullough.

The model's reasoning: These claims match the paper's findings on PELP1-mediated BCAS3 expression and BCAS3's acetyltransferase and ER-α coactivator activities.

The sentence the footnote ends: In addition BCAS3 possesses histone acetyltransferase activity and itself appears to act as a coactivator of ER-α.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-17243 / 15349652 / #1, read from revision 1301002514 of BCAS3.

The model's reasoning: The PELP1 requirement and resulting positive-feedback mechanism directly repeat the retracted paper's findings and conclusion.

The sentence the footnote ends: Hence BCAS3 apparently is involved in a positive feedback loop leading to ER-α mediated signal amplification.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-17243 / 15349652 / #2, read from revision 1301002514 of BCAS3.

The model's reasoning: The claim about related processes incorporates the retracted paper's finding that LAPF/Phafin1 induces apoptosis through a lysosomal pathway; the other citation plausibly supports related activity of Phafin2.

Other citations on it:

• Matsuda-Lennikov M, Suizu F, Hirata N, Hashimoto M, Kimura K, Nagamine T, et al. (2014-01-08). Chiorini JA (ed.). "Lysosomal interaction of Akt with Phafin2: a critical step in the induction of autophagy". PLOS ONE. 9 (1) e79795. Bibcode:2014PLoSO...979795M. doi:10.1371/journal.pone.0079795. PMC 3885392. PMID 24416124.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.94. Finding RW-29253 / 67069486 / #1, read from revision 1314999644 of Pib2.

The model's reasoning: The paragraph describes the paper's subject and explicitly identifies its retraction rather than presenting its finding as established fact.

The sentence the footnote ends: In December 2020, the journal retracted the aforementioned article, contrary to the COPE guidelines, which advise that the original paper should remain available but prominently marked as retracted.

Where the model says the article warns about this paper:

• In the same section (Liver disease inquiries): In December 2020, the journal retracted the aforementioned article, contrary to the COPE guidelines, which advise that the original paper should remain available but prominently marked as retracted.
• In the same section (Liver disease inquiries): According to one of the co-authors of the paper, this retraction happened because Herbalife engaged DSK Legal, a New Delhi–based law firm, to consistently issue legal threats to the journal's editor-in-chief.

Other citations on it:

• Bik, Elisabeth (December 20, 2020). "Paper about Herbalife®-related patient death removed after company threatens to sue the journal". Science Integrity Digest. Archived from the original on July 31, 2024. Retrieved December 21, 2020.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-23577 / 753338 / #1, read from revision 1373917333 of Herbalife.

The model's reasoning: This reports a call for retractions rather than relying on the paper's therapeutic findings; the accompanying investigation and misconduct sources plausibly support it.

Where the model says the article warns about this paper:

• In the same section (Efficacy): However, it was revealed that Eysenck (who died in 1997) falsified data in his studies about this subject, fabricating data that would indicate that behavioral therapy enables achievements that are impossible to believe.
• In the same section (Efficacy): Fourteen of his papers were retracted by journals in 2020, and journals issued 64 statements of concern about publications by him.
• In the same section (Efficacy): Rod Buchanan, a biographer of Eysenck, has argued that 87 publications by Eysenck should be retracted.

Other citations on it:

• King's College London (May 2019). King's College London enquiry into publications authored by Professor Hans Eysenck with Professor Ronald Grossarth-Maticek. https://retractionwatch.com/wp-content/uploads/2019/10/HE-Enquiry.pdf
• Marks, D. F. (2019). The Hans Eysenck affair: Time to correct the scientific record. Journal of Health Psychology, 1–12.
• Marks, D. F. & Buchanan, R. D. (2020). King's College London's enquiry into Hans J. Eysenck's 'unsafe' publications must be properly completed. Journal of Health Psychology, 25, 1, 3–6.
• O'Grady, C. (15 July 2020). Misconduct allegations push psychology hero off his pedestal. Science Mag. https://www.science.org/content/article/misconduct-allegations-push-psychology-hero-his-pedestal
• Pelosi, A. J. (2019). Personality and fatal diseases: Revisiting a scientific scandal. Journal of Health Psychology, 24, 4, 421–439.
• Rolls, G. (2015). Classic case studies in psychology. Routledge Taylor & Francis.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-29957 / 24931 / #1, read from revision 1372436934 of Psychotherapy.

The model's reasoning: The paragraph recounts the paper's claims as part of its history and explicitly identifies the paper as fraudulent and retracted, rather than presenting its findings as valid.

Where the model says the article warns about this paper:

• In the same section (False claims about autism): False claims about autism
• In the same section (False claims about autism): In 1998 Andrew Wakefield et al. published a fraudulent paper about twelve children, reportedly with bowel symptoms and autism or other disorders acquired soon after administration of MMR vaccine, while supporting a competing vaccine.
• In the same section (False claims about autism): In 2010, Wakefield's research was found by the General Medical Council to have been "dishonest", and The Lancet fully retracted the paper.
• In the same section (False claims about autism): Three months following The Lancet's retraction, Wakefield was struck off the UK medical register, with a statement identifying deliberate falsification in the research published in The Lancet, and was barred from practising medicine in the UK.
• In the same section (False claims about autism): The research was declared fraudulent in 2011 by the British Medical Journal.
• In the same section (False claims about autism): This has had no effect on autism rates in the country, further disproving the MMR autism hypothesis.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-4036 / 69149 / #1, read from revision 1375536739 of MMR vaccine.

The model's reasoning: The paragraph reports the paper's genetic-relatedness conclusion without saying locally that it was withdrawn; the withdrawal warning appears in the next paragraph.

The sentence the footnote ends: In 2001, Antonio Arnaiz-Villena co-authored a research paper in the journal on the genetic history of Jews, Palestinians, and other Mediterranean populations.

Where the model says the article warns about this paper:

• In the same section (Controversies): Following a large number of complaints, the article was withdrawn from the journal and deleted from the scientific archive.
• In the same section (Controversies): Academics who had already received a copy of the journal were urged to "physically remove" the article pages in a move that had no precedent in research publishing.
• In the same section (Controversies): Andrew Goffey, a senior lecturer at Middlesex University in England, observed that "it was conceded that the article had not been removed on the basis of its scientific evidence," failed to find anything offensive in the paper.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-3030 / 25579891 / #1, read from revision 1362165677 of Human Immunology.

The model's reasoning: This describes the paper's publication and retraction, not its stem-cell findings; the accompanying news report and retraction notice plausibly support this account.

Where the model says the article warns about this paper:

• In the same section (Retractions): In 2024, a paper titled "Pluripotency of Mesenchymal Stem Cells Derived from Adult Marrow", published in 2002, was retracted due to concerns raised regarding some of the panels shown in a figure, making it the most-cited retracted paper ever.

Other citations on it:

• "University of Minnesota retracts pioneering studies in stem cells, Alzheimer's disease". 7 July 2024. Archived from the original on 7 July 2024. Retrieved 7 July 2024.
• Jiang, Yuehua; Jahagirdar, Balkrishna N.; Reinhardt, R. Lee; Schwartz, Robert E.; Keene, C. Dirk; Ortiz-Gonzalez, Xilma R.; Reyes, Morayma; Lenvik, Todd; Lund, Troy; Blackstad, Mark; Du, Jingbo; Aldrich, Sara; Lisberg, Aaron; Low, Walter C.; Largaespada, David A. (June 2024). "Retraction Note: Pluripotency of mesenchymal stem cells derived from adult marrow". Nature. 630 (8018): 1020. Bibcode:2024Natur.630.1020J. doi:10.1038/s41586-024-07653-0. ISSN 1476-4687. PMID 38886620.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-54180 / 43427 / #1, read from revision 1378356736 of Nature (journal).

The model's reasoning: The paragraph describes the bogus paper and its retraction, without relying on its purported findings.

Where the model says the article warns about this paper:

• In the same section (Sokal III): In October 2021, the scholarly journal Higher Education Quarterly published a bogus article "authored" by "Sage Owens" and "Kal Avers-Lynde III".
• In the same section (Sokal III): The Quarterly retracted the article.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-34471 / 28547 / #1, read from revision 1365391661 of Sokal affair.

The model's reasoning: This recounts the paper's claimed stem-cell result in a research-misconduct paragraph that explicitly identifies the work's subsequent retraction.

Where the model says the article warns about this paper:

• In the same section (Whistleblowing in the scientific community): Independent committees, as well as journalists, scrutinized the research data and methodology, leading to an eventual retraction of his work.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-1007 / 33821 / #1, read from revision 1377556143 of Whistleblowing.

The model's reasoning: These attributed inhibition and lifespan-extension claims repeat this paper’s findings without a local retraction warning; an editor should check whether this paragraph has independent sourcing.

Where the model says the article warns about this paper:

• In the opening paragraphs: However, the entire work behind the discovery of this compound has since been found to be falsified and the authors of the original reports have retracted all their claims.
• In the opening paragraphs: The original report garnered scientific attention, but was retracted in 2008. The retraction states that the screen to identify CGK733 as an anti-senescence agent was not carried out; experiments exploring the cellular effects of CGK733 were misrepresented; the identification of ATM as the target of CGK733 was fabricated; a compound which was essential for ATM target validation had not been synthesized; and the chemical structure of CGK733 was misrepresented.

The model's confidence: 0.97. Finding RW-3330 / 8729837 / #2, read from revision 1357177569 of CGK733 fraud.

The model's reasoning: The caption repeats the retracted paper's claimed synthesis without a local retraction warning; an editor should check whether the caption has independent sourcing.

Where the model says the article warns about this paper:

• In the same section (Controversy over structure): The scientific community then began criticizing La Clair's work, claiming that his work was sloppy or that he fabricated data.
• In the same section (Controversy over structure): La Clair's publication of his purported synthesis was retracted in 2012, citing a lack of validation of its claims.
• In the same section (Controversy over structure): In response to the controversy, Nobel Prize-winning synthetic chemist E.J. Corey remarked, "Occasionally, blatantly wrong science is published, and to the credit of synthetic chemistry, the corrections usually come quickly and cleanly."

The model's confidence: 0.88. Finding RW-779 / 29839319 / #2, read from revision 1292433851 of Hexacyclinol.

The model's reasoning: This describes a mutant axon-guidance phenotype, not the retracted paper’s finding that ligand-triggered UNC5–DCC association converts attraction to repulsion; the cited C. elegans genetics paper plausibly supports this background claim.

Other citations on it:

• Neuroglia and pioneer neurons express UNC-6 to provide global and local netrin cues for guiding migrations in C. elegans
• Serafini T, Kennedy TE, Galko MJ, Mirzayan C, Jessell TM, Tessier-Lavigne M (August 1994). "The netrins define a family of axon outgrowth-promoting proteins homologous to C. elegans UNC-6". Cell. 78 (3): 409–24. doi:10.1016/0092-8674(94)90420-0. PMID 8062384. S2CID 22666205.
• Hedgecock EM, Culotti JG, Hall DH (January 1990). "The unc-5, unc-6, and unc-40 genes guide circumferential migrations of pioneer axons and mesodermal cells on the epidermis in C. elegans". Neuron. 4 (1): 61–85. doi:10.1016/0896-6273(90)90444-K. PMID 2310575. S2CID 23974242.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.88. Finding RW-49334 / 958 / #1, read from revision 1368873164 of Axon.

The model's reasoning: The paragraph reports the paper's conclusion as incorrect in the context of editorial resignations, rather than relying on it as evidence.

The sentence the footnote ends: In 2021, five members of the editorial board of the journal Vaccines resigned after Vaccines published a controversial article that misused data to reach the incorrect conclusion that vaccines against COVID-19 had no clear benefit.

Where the model says the article warns about this paper:

• In the same section (Resignations of editors): a controversial article that misused data to reach the incorrect conclusion that vaccines against COVID-19 had no clear benefit.
• In the same section (Controversial articles): The article was heavily criticized for misusing data and as a result reaching a false conclusion.
• In the same section (Controversial articles): Shortly afterwards, the journal published an "expression of concern" regarding the article and opened an investigation into the review process.
• In the same section (Controversial articles): Vaccines eventually decided to retract the article.

Other citations on it:

• Wadman, Meredith (1 July 2021). "Scientists quit journal board, protesting 'grossly irresponsible' study claiming COVID-19 vaccines kill". Science. doi:10.1126/science.abk2629. S2CID 243092907. Retrieved 2 July 2021.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-30287 / 13943035 / #1, read from revision 1378385916 of MDPI.

The model's reasoning: The same paragraph explicitly identifies the conclusion as false and states that the article was retracted; the citation documents the controversy.

Where the model says the article warns about this paper:

• In the same section (Resignations of editors): a controversial article that misused data to reach the incorrect conclusion that vaccines against COVID-19 had no clear benefit.
• In the same section (Controversial articles): The article was heavily criticized for misusing data and as a result reaching a false conclusion.
• In the same section (Controversial articles): Shortly afterwards, the journal published an "expression of concern" regarding the article and opened an investigation into the review process.
• In the same section (Controversial articles): Vaccines eventually decided to retract the article.

Other citations on it:

• Ansede, Manuel; Urra, Susana (22 July 2021). "Harald Walach, the researcher contributing misinformation about the Covid-19 pandemic". EL PAÍS.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-30287 / 13943035 / #2, read from revision 1378385916 of MDPI.

The model's reasoning: This describes reported image fraud in the paper, not its scientific findings; the accompanying blog citation plausibly supports the report.

Where the model says the article warns about this paper:

• In the same section (Photo manipulation): One Nature paper published in 2009 has subsequently been reported to contain around 20 separate instances of image fraud.
• Later (Journals): In one case, Nature allowed a corrigendum to be published despite clear evidence of image fraud. Subsequent retraction of the paper required the actions of an independent whistleblower.

Other citations on it:

• 11jigen (2012-01-15). "Shigeaki Kato (the University of Tokyo): DNA demethylation in hormone-induced transcriptional derepression". Katolab-imagefraud.blogspot.co.uk. Retrieved 2013-08-04.{{cite web}}: CS1 maint: numeric names: authors list (link)

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-2240 / 29537 / #1, read from revision 1378579895 of Scientific misconduct.

The model's reasoning: This concerns the journal's handling of the paper rather than its findings, and the paragraph explicitly states that the paper was subsequently retracted.

Where the model says the article warns about this paper:

• Earlier (Photo manipulation): One Nature paper published in 2009 has subsequently been reported to contain around 20 separate instances of image fraud.
• In the same section (Journals): In one case, Nature allowed a corrigendum to be published despite clear evidence of image fraud. Subsequent retraction of the paper required the actions of an independent whistleblower.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-2240 / 29537 / #2, read from revision 1378579895 of Scientific misconduct.

The model's reasoning: This describes an immune mechanism in the study's vaccine context, but the abstract does not establish whether the paper demonstrated it or merely discussed background immunology.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.85. Finding RW-5206 / 68892222 / #2, read from revision 1353163515 of Bacteriophage AP205.

The model's reasoning: The abstract does not establish whether this RNA-mediated mechanism was a finding of the retracted study or background drawn from other research.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.85. Finding RW-5206 / 68892222 / #3, read from revision 1353163515 of Bacteriophage AP205.

The model's reasoning: This is a general regulatory comparison, rather than the paper's survey finding about public reception; it needs an unretracted background source.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.76. Finding RW-889 / 12339 / #1, read from revision 1374510745 of Genetically modified organism.

The model's reasoning: The paragraph reports the study's arsenic-DNA claim but explicitly describes failed replication and the paper's retraction, rather than presenting the finding as established.

Where the model says the article warns about this paper:

• In the same section (Arsenic as an alternative to phosphorus): This claim was heavily criticized almost immediately after publication for the perceived lack of appropriate controls.
• In the same section (Arsenic as an alternative to phosphorus): Other authors were unable to reproduce their results and showed that the study had issues with phosphate contamination, suggesting that the low amounts present could sustain extremophile lifeforms.
• In the same section (Arsenic as an alternative to phosphorus): The 2010 paper was retracted in 2025.

Other citations on it:

• "NASA-Funded Research Discovers Life Built With Toxic Chemical". NASA.gov. 2 December 2010. Archived from the original on 2011-08-28. Retrieved 2010-12-02.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-66335 / 7316 / #1, read from revision 1376527071 of Hypothetical types of biochemistry.

The model's reasoning: This proposed mechanism is reported within a paragraph that explicitly identifies the study's failed replication and retraction.

Where the model says the article warns about this paper:

• In the same section (Arsenic as an alternative to phosphorus): This claim was heavily criticized almost immediately after publication for the perceived lack of appropriate controls.
• In the same section (Arsenic as an alternative to phosphorus): Other authors were unable to reproduce their results and showed that the study had issues with phosphate contamination, suggesting that the low amounts present could sustain extremophile lifeforms.
• In the same section (Arsenic as an alternative to phosphorus): The 2010 paper was retracted in 2025.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-66335 / 7316 / #2, read from revision 1376527071 of Hypothetical types of biochemistry.

The model's reasoning: The paragraph reports the original description but immediately explains that the species was invalidated because its holotype was a forgery, rather than endorsing the finding.

Where the model says the article warns about this paper:

• In the same section (Taxonomy): In 2012, "A. kurteni" was invalidated as a species when the holotype was determined to be a forgery composed of Miocene-aged fragments.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-5686 / 69508 / #1, read from revision 1378485511 of Acinonyx.

The model's reasoning: This describes the paper's retraction rather than relying on its clinical findings; both news sources plausibly support the retraction account.

Where the model says the article warns about this paper:

• In the same section (Surgisphere study on the use of hydroxychloroquine and chloroquine (2020)): The authors of the study then asked The Lancet to retract the article, which was done on 3 June 2020.
• In the same section (Surgisphere study on the use of hydroxychloroquine and chloroquine (2020)): To improve quality control, the editors of The Lancet Group announced changes to the editorial policy in a comment titled "Learning from a retraction" which was published on 22 September 2020.

Other citations on it:

• Hopkins, Jared (5 June 2020). "Hydroxychloroquine Studies Tied to Data Firm Surgisphere Retracted". The Wall Street Journal. Archived from the original on 5 January 2022. Retrieved 31 December 2021.
• "Covid-19: Lancet retracts paper that halted hydroxychloroquine trials". the Guardian. 4 June 2020. Archived from the original on 17 August 2022. Retrieved 31 December 2021.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-23529 / 463899 / #2, read from revision 1372163734 of The Lancet.

The model's reasoning: This repeats the retracted study's diagnostic-performance claim without a retraction warning in this paragraph; an editor should check whether any local sources independently support it.

Where the model found it: A study was initiated which suggested that the EPCA-2 protein serum assay exhibits favorable performance characteristics which are potentially superior to serum PSA. However more studies are necessary to see if this test will retain its sensitivity when used in a screening population.

Where the model says the article warns about this paper:

• In the same section (EPCA-2 versus PSA): This paper has since been retracted by the publisher.
• In the same section (EPCA-2 versus PSA): In September 2008 the industry sponsor of EPCA-2, Onconome sued Dr Robert Getzenberg, JHU, and the University of Pittsburgh, his previous institution, claiming that Getzenberg misrepresented and falsified data related to EPCA-2 after Onconome sponsored 13 million dollars of research over five years in Getzenberg's labs at University of Pittsburgh and Johns Hopkins for a blood test for prostate cancer.

The model's confidence: 0.98. Finding RW-1844 / 6928455 / #3, read from revision 1367468495 of Early prostate cancer antigen-2.

The model's reasoning: This reports the stem-cell treatment claim historically, and the same paragraph explicitly notes failed replication and retractions for scientific fraud.

The sentence the footnote ends: In the 2000s, the discovery of adult endogenous cardiac stem cells was reported, and studies were published that claimed that various stem cell lineages, including bone marrow stem cells were able to differentiate into cardiomyocytes, and could be used to treat heart failure.

Where the model says the article warns about this paper:

• In the same section (Regeneration): However, other teams were unable to replicate these findings, and many of the original studies were later retracted for scientific fraud.

Other citations on it:

• Orlic D, Kajstura J, Chimenti S, Jakoniuk I, Anderson SM, Li B, Pickel K, McKay R, Nadal-Ginard B, Bodine DM, Anversa P (April 2001). "Bone marrow cells regenerate infarcted myocardium". Nature. 410 (6829): 701–705. Bibcode:2001Natur.410..701O. doi:10.1038/35070587. PMID 11287958. S2CID 4424399.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.98. Finding RW-20119 / 424348 / #1, read from revision 1371774857 of Cardiac muscle.

The model's reasoning: The paragraph recounts the paper's conclusion but explicitly explains the species-identification error and subsequent retraction, rather than presenting the finding as valid.

Where the model says the article warns about this paper:

• In the same section (Contested wildlife mutation research): The authors discovered they had incorrectly classified the species of vole and were genetically comparing two different vole species. They issued a retraction in 1997.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-2098 / 4155456 / #1, read from revision 1376665179 of Effects of the Chernobyl disaster.

The model's reasoning: This records publication rather than relying on the study's results; the same paragraph explicitly reports its retraction.

Where the model says the article warns about this paper:

• In the same section (COVID-19): The study was retracted on 17 December 2024 over a lack of informed consent for patients and after scientists—including three of its co-authors—raised concerns over methodological problems.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-59596 / 27015334 / #1, read from revision 1372760393 of Didier Raoult.

The model's reasoning: The description of trial success refers to this paper's finding, but appears as background to a scientific-integrity controversy rather than evidence of efficacy; an editor should check the paragraph's own sources for this characterization.

Where the model found it: On 5 May 2021, Elisabeth Bik (who specializes in identifying manipulated images in scientific papers) raised concerns about dozens of Raoult's papers—including ethical, procedural, and methodological problems in a March 2020 paper reporting success in a small hydroxychloroquine trial. Raoult's lawyer subsequently announced that Raoult was accusing and suing the scientific integrity consultant of harassment and blackmail. The French non-profit association Citizen4Science, formed by scientists and citizens, published a press release and a petition that day, denouncing the harassment of scientists and defenders of science integrity, specifically defending Bik and calling on French authorities to intervene and for journalists to look into the matter. Several French newspapers immediately reported Citizen4Science's initiative. The petition was signed by thousands of scientists and others throughout the world. By 22 May 2021 Raoult had begun legal proceedings against Bik. There followed various articles in international mainstream media supporting Bik, and an article in Science updated on 4 June 2021 in issue 6546, reporting over 3,000 signatures for the Citizen4Science petition.

Where the model says the article warns about this paper:

• In the same section (COVID-19): The study was retracted on 17 December 2024 over a lack of informed consent for patients and after scientists—including three of its co-authors—raised concerns over methodological problems.

The model's confidence: 0.94. Finding RW-59596 / 27015334 / #2, read from revision 1372760393 of Didier Raoult.

The model's reasoning: The review evaluated saline irrigation for chronic rhinosinusitis, not steam inhalation; the accompanying patient guidance plausibly supports this background treatment claim.

Other citations on it:

• Consumer Reports, American Academy of Allergy, Asthma, and Immunology (July 2012), "Treating sinusitis: Don't rush to antibiotics" (PDF), Choosing Wisely: an initiative of the ABIM Foundation, Consumer Reports, archived (PDF) from the original on January 24, 2013, retrieved August 14, 2012

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.97. Finding RW-31614 / 28598 / #1, read from revision 1373933225 of Sinusitis.

The model's reasoning: This is background comparative anatomy, not the paper's proposed conclusions about metamorphosis, heterochrony, and evolutionary diversity; the publisher-error retraction calls for checking the replacement source.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.88. Finding RW-10920 / 20975731 / #1, read from revision 1371880440 of Lamprey.

The model's reasoning: This is a general classification of supportive care, not a finding of the review about herbal medicines; the withdrawn review should be replaced with a current source for this background claim.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.98. Finding RW-31408 / 452461 / #1, read from revision 1376498073 of Myocarditis.

The model's reasoning: This describes the paper and its historical consequences while explicitly identifying it as retracted and fraudulent, rather than endorsing its finding.

Where the model says the article warns about this paper:

• In the same section (False claims about autism): False claims about autism
• In the same section (False claims about autism): In 1998, Andrew Wakefield et al. published a now retracted and fraudulent The Lancet paper linking the MMR vaccine to autism, leading to a decline in vaccination rates.
• In the same section (False claims about autism): In Japan, where the MMR vaccine is not used as a combined vaccine, autism rates have remained unaffected, further disproving Wakefield's hypothesis.
• In the same section (False claims about autism): Wakefield's disproven theory that vaccines cause autism

Other citations on it:

• The Editors of The Lancet (February 2010). "Retraction—Ileal-lymphoid-nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children". Lancet. 375 (9713): 445. doi:10.1016/S0140-6736(10)60175-4. PMID 20137807. S2CID 26364726.
• Triggle N (2 February 2010). "Lancet accepts MMR study 'false'". BBC News. Archived from the original on 3 November 2021. Retrieved 11 June 2022.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-4036 / 21053550 / #1, read from revision 1377781515 of Measles vaccine.

The model's reasoning: This describes the paper's publication and retraction in a hoax controversy, without asserting its findings.

Where the model says the article warns about this paper:

• In the same section (Academic practices): In 2018, another Grievance studies affair article "Human reactions to rape culture and queer performativity at urban dog parks in Portland, Oregon" was published in Gender, Place & Culture, which was also retracted later that year.

Other citations on it:

• Pluckrose, Helen; Lindsay, James A.; Boghossian, Peter (2 October 2018). "Academic Grievance Studies and the Corruption of Scholarship". Areo Magazine. Archived from the original on 10 October 2018.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-19280 / 558559 / #1, read from revision 1370320390 of Taylor & Francis.

The model's reasoning: This reports the paper's efficacy finding, but the same paragraph explicitly tells readers that the study was retracted.

Where the model says the article warns about this paper:

• In the same section (Research): However, in 2015, the journal retracted the study, claiming that the researchers had acted unethically when providing Chinese children golden rice without their parents' consent.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-948 / 246458 / #1, read from revision 1374816022 of Golden rice.

The model's reasoning: This reports the paper's ethics statement rather than its scientific finding, and the paragraph explains the subsequent retraction for lack of parental consent.

Where the model says the article warns about this paper:

• In the same section (Research): However, in 2015, the journal retracted the study, claiming that the researchers had acted unethically when providing Chinese children golden rice without their parents' consent.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-948 / 246458 / #2, read from revision 1374816022 of Golden rice.

The model's reasoning: The citation identifies the paper in an account of its research-ethics controversy, without relying on its efficacy findings.

Where the model says the article warns about this paper:

• Earlier (Research): However, in 2015, the journal retracted the study, claiming that the researchers had acted unethically when providing Chinese children golden rice without their parents' consent.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-948 / 246458 / #3, read from revision 1374816022 of Golden rice.

The model's reasoning: The sentence reports the study's claims as part of its retraction history, explicitly identifying it as retracted rather than presenting its findings as valid.

Where the model says the article warns about this paper:

• In the same section (Cameroon): A since-retracted 2018 study
• In the same section (Cameroon): the study had been reviewed and found to be "scientifically worthless, contradictory, and in part ethically problematic"
• In the same section (Cameroon): In August 2019 the study was retracted by the journal which had published it because the editors concluded after a complaint and investigation revealed that the study had never actually been conducted.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-20991 / 25572706 / #1, read from revision 1344445606 of Miracle Mineral Supplement.

The model's reasoning: The item describes the paper's claim as fraudulent and explicitly reports its retraction rather than presenting its findings as true.

The whole list item: Andrew Wakefield (UK), a former surgeon and senior lecturer at the Royal Free Hospital in London, was found guilty of dishonesty in his research on the MMR vaccine and banned from medicine by the UK General Medical Council following an investigation by Brian Deer of the Sunday Times. Wakefield's claims of a link between the MMR vaccine, autism and inflammatory bowel disease have been reported in the British Medical Journal as "based not on bad science but on a deliberate fraud", and the 1998 paper originally presenting his theory was retracted in 2010 by The Lancet. Wakefield was unsuccessful in an attempt to sue detractors/critics for libel and defamation. Wakefield has had two papers retracted and one corrected.

Where the model says the article warns about this paper:

• In the same section (Biology and biomedical sciences): Wakefield's claims of a link between the MMR vaccine, autism and inflammatory bowel disease have been reported in the British Medical Journal as "based not on bad science but on a deliberate fraud", and the 1998 paper originally presenting his theory was retracted in 2010 by The Lancet.

Other citations on it:

• "Dr. Andrew Jeremy Wakefield: Determination on Serious Professional Misconduct (SPM) and Sanction" (PDF). General Medical Council. 24 May 2010. Archived from the original (PDF) on 9 August 2011. Retrieved 10 August 2011.
• Godlee, F. (6 January 2011). "The fraud behind the MMR scare". BMJ. 342: d22. doi:10.1136/bmj.d22. S2CID 73020733. Retrieved 30 March 2018.
• Dyer, C. (2 February 2010). "Lancet retracts Wakefield's MMR paper". British Medical Journal. 340: c696. doi:10.1136/bmj.c696. PMID 20124366. S2CID 43465004. Retrieved 30 March 2018.
• "Wakefield Libel Suit Rejected". MedPage Today LLC. 6 August 2012. Retrieved 30 March 2018.
• Dyer, C. (6 August 2012). "Texas judge throws out Wakefield's libel action against BMJ". British Medical Journal. 345 e5328. doi:10.1136/bmj.e5328. PMID 22867918. Retrieved 30 March 2018.
• "Retraction Watch Database – Andrew Wakefield". Retraction Watch. The Center for Scientific Integrity. Retrieved 22 January 2019.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-4036 / 56023027 / #1, read from revision 1374378694 of List of scientific misconduct incidents.

The model's reasoning: The item reports the study's proposal as part of its misconduct history and explicitly states that the same paper was retracted for manipulated images.

The whole list item: Sylvain Lesné (France, US), a neuroscientist, co-authored a study in Nature proposing that a specific amyloid-beta protein assembly, Aβ*56, impairs memory and contributes to Alzheimer's disease pathogenesis. In 2024, Nature retracted this paper after investigations revealed manipulated images, with co-authors Michela Gallagher and Karen H. Ashe agreeing to the retraction, while Lesné disagreed. The story was revealed in 2022 by Charles Piller in Science, based upon investigations by Matthew Schrag, a neuroscientist and physician at Vanderbilt University. This narrative is a key component of Piller's 2025 book Doctored: Fraud, Arrogance, and Tragedy in the Quest to Cure Alzheimer's.

Where the model says the article warns about this paper:

• In the same section (Biology and biomedical sciences): In 2024, Nature retracted this paper after investigations revealed manipulated images, with co-authors Michela Gallagher and Karen H. Ashe agreeing to the retraction, while Lesné disagreed.

Other citations on it:

• Lesné, Sylvain; Koh, Ming Teng; Kotilinek, Linda; Kayed, Rakez; Glabe, Charles G.; Yang, Austin; Gallagher, Michela; Ashe, Karen H. (July 2024). "Retraction Note: A specific amyloid-β protein assembly in the brain impairs memory". Nature. 631 (8019): 240. Bibcode:2024Natur.631..240L. doi:10.1038/s41586-024-07691-8. ISSN 1476-4687. PMID 38914864.
• "Potential fabrication in research images threatens key theory of Alzheimer's disease". science.org. Retrieved 30 January 2025.
• Piller, Charles (13 February 2025). Doctored: Fraud, Arrogance and Tragedy in the Quest to Cure Alzheimers. Icon Books. ISBN 978-1-83773-259-3.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-54364 / 56023027 / #1, read from revision 1374378694 of List of scientific misconduct incidents.

The model's reasoning: The item describes retractions and compromised editorial handling, not the paper's findings about priming and mineralisation in soils.

The whole list item: Yakov Kuzyakov lost 4 papers because he edited his own papers in the Elsevier journal European Journal of Soil Biology. For three of these papers one of the reviewers was also closely linked to Y. Kuzyakov. He lost a fourth paper in a second Elsevier journal, Science of the Total Environment, due to suspicious changes in authors on the paper during the review process.

Where the model says the article warns about this paper:

• In the same section (Biology and biomedical sciences): Yakov Kuzyakov lost 4 papers because he edited his own papers in the Elsevier journal European Journal of Soil Biology. For three of these papers one of the reviewers was also closely linked to Y. Kuzyakov.

Other citations on it:

• Liang, Qiong; Chen, Haiqing; Gong, Yuanshi; Yang, Hefa; Fan, Mingsheng; Kuzyakov, Yakov (1 January 2014). "RETRACTED: Effects of 15 years of manure and mineral fertilizers on enzyme activities in particle-size fractions in a North China Plain soil". European Journal of Soil Biology. 60: 112–119. Bibcode:2014EJSB...60..112L. doi:10.1016/j.ejsobi.2013.11.009.
• Maharjan, Menuka; Maranguit, Deejay; Kuzyakov, Yakov (May 2018). "RETRACTED: Phosphorus fractions in subtropical soils depending on land use". European Journal of Soil Biology. 87: 17–24. Bibcode:2018EJSB...87...17M. doi:10.1016/j.ejsobi.2018.04.002.
• Qiu, Husen; Ge, Tida; Liu, Jieyun; Chen, Xiangbi; Hu, Yajun; Wu, Jinshui; Su, Yirong; Kuzyakov, Yakov (January 2018). "RETRACTED: Effects of biotic and abiotic factors on soil organic matter mineralization: Experiments and structural modeling analysis". European Journal of Soil Biology. 84: 27–34. Bibcode:2018EJSB...84...27Q. doi:10.1016/j.ejsobi.2017.12.003.
• Wu, Cuiyan; Huang, Wei; Liu, Yixian; Li, Han; Ding, Shuai; Zhu, Zhenke; Wang, Feng; Dijkstra, Feike A.; Zhang, Guangbin; Kuzyakov, Yakov; Cheng, Weiguo; Xiao, Mouliang; Ge, Tida (10 December 2024). "RETRACTED: Physicochemically protected organic carbon release is the rate-limiting step of rhizosphere priming in paddy soils". Science of the Total Environment. 955 176859. Bibcode:2024ScTEn.95576859W. doi:10.1016/j.scitotenv.2024.176859. PMID 39414056.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-73244 / 56023027 / #1, read from revision 1379120205 of List of scientific misconduct incidents.

The model's reasoning: The item describes retractions and compromised editorial review, not the paper's findings about soil organic matter mineralization.

The whole list item: Yakov Kuzyakov lost 4 papers because he edited his own papers in the Elsevier journal European Journal of Soil Biology. For three of these papers one of the reviewers was also closely linked to Y. Kuzyakov. He lost a fourth paper in a second Elsevier journal, Science of the Total Environment, due to suspicious changes in authors on the paper during the review process.

Where the model says the article warns about this paper:

• In the same section (Biology and biomedical sciences): Yakov Kuzyakov lost 4 papers because he edited his own papers in the Elsevier journal European Journal of Soil Biology. For three of these papers one of the reviewers was also closely linked to Y. Kuzyakov.

Other citations on it:

• Liang, Qiong; Chen, Haiqing; Gong, Yuanshi; Yang, Hefa; Fan, Mingsheng; Kuzyakov, Yakov (1 January 2014). "RETRACTED: Effects of 15 years of manure and mineral fertilizers on enzyme activities in particle-size fractions in a North China Plain soil". European Journal of Soil Biology. 60: 112–119. Bibcode:2014EJSB...60..112L. doi:10.1016/j.ejsobi.2013.11.009.
• Maharjan, Menuka; Maranguit, Deejay; Kuzyakov, Yakov (May 2018). "RETRACTED: Phosphorus fractions in subtropical soils depending on land use". European Journal of Soil Biology. 87: 17–24. Bibcode:2018EJSB...87...17M. doi:10.1016/j.ejsobi.2018.04.002.
• Qiu, Husen; Zheng, Xiaodong; Ge, Tida; Dorodnikov, Maxim; Chen, Xiangbi; Hu, Yajun; Kuzyakov, Yakov; Wu, Jinshui; Su, Yirong; Zhang, Zhenhua (November 2017). "RETRACTED: Weaker priming and mineralisation of low molecular weight organic substances in paddy than in upland soil". European Journal of Soil Biology. 83: 9–17. Bibcode:2017EJSB...83....9Q. doi:10.1016/j.ejsobi.2017.09.008.
• Wu, Cuiyan; Huang, Wei; Liu, Yixian; Li, Han; Ding, Shuai; Zhu, Zhenke; Wang, Feng; Dijkstra, Feike A.; Zhang, Guangbin; Kuzyakov, Yakov; Cheng, Weiguo; Xiao, Mouliang; Ge, Tida (10 December 2024). "RETRACTED: Physicochemically protected organic carbon release is the rate-limiting step of rhizosphere priming in paddy soils". Science of the Total Environment. 955 176859. Bibcode:2024ScTEn.95576859W. doi:10.1016/j.scitotenv.2024.176859. PMID 39414056.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-73246 / 56023027 / #1, read from revision 1379120205 of List of scientific misconduct incidents.

The model's reasoning: This reports retractions and editorial misconduct, not the paper's findings about soil phosphorus fractions.

The whole list item: Yakov Kuzyakov lost 4 papers because he edited his own papers in the Elsevier journal European Journal of Soil Biology. For three of these papers one of the reviewers was also closely linked to Y. Kuzyakov. He lost a fourth paper in a second Elsevier journal, Science of the Total Environment, due to suspicious changes in authors on the paper during the review process.

Where the model says the article warns about this paper:

• In the same section (Biology and biomedical sciences): Yakov Kuzyakov lost 4 papers because he edited his own papers in the Elsevier journal European Journal of Soil Biology. For three of these papers one of the reviewers was also closely linked to Y. Kuzyakov.

Other citations on it:

• Liang, Qiong; Chen, Haiqing; Gong, Yuanshi; Yang, Hefa; Fan, Mingsheng; Kuzyakov, Yakov (1 January 2014). "RETRACTED: Effects of 15 years of manure and mineral fertilizers on enzyme activities in particle-size fractions in a North China Plain soil". European Journal of Soil Biology. 60: 112–119. Bibcode:2014EJSB...60..112L. doi:10.1016/j.ejsobi.2013.11.009.
• Qiu, Husen; Zheng, Xiaodong; Ge, Tida; Dorodnikov, Maxim; Chen, Xiangbi; Hu, Yajun; Kuzyakov, Yakov; Wu, Jinshui; Su, Yirong; Zhang, Zhenhua (November 2017). "RETRACTED: Weaker priming and mineralisation of low molecular weight organic substances in paddy than in upland soil". European Journal of Soil Biology. 83: 9–17. Bibcode:2017EJSB...83....9Q. doi:10.1016/j.ejsobi.2017.09.008.
• Qiu, Husen; Ge, Tida; Liu, Jieyun; Chen, Xiangbi; Hu, Yajun; Wu, Jinshui; Su, Yirong; Kuzyakov, Yakov (January 2018). "RETRACTED: Effects of biotic and abiotic factors on soil organic matter mineralization: Experiments and structural modeling analysis". European Journal of Soil Biology. 84: 27–34. Bibcode:2018EJSB...84...27Q. doi:10.1016/j.ejsobi.2017.12.003.
• Wu, Cuiyan; Huang, Wei; Liu, Yixian; Li, Han; Ding, Shuai; Zhu, Zhenke; Wang, Feng; Dijkstra, Feike A.; Zhang, Guangbin; Kuzyakov, Yakov; Cheng, Weiguo; Xiao, Mouliang; Ge, Tida (10 December 2024). "RETRACTED: Physicochemically protected organic carbon release is the rate-limiting step of rhizosphere priming in paddy soils". Science of the Total Environment. 955 176859. Bibcode:2024ScTEn.95576859W. doi:10.1016/j.scitotenv.2024.176859. PMID 39414056.

The reference isn't tagged {{Retracted}}.

The model's confidence: 0.99. Finding RW-73247 / 56023027 / #1, read from revision 1379120205 of List of scientific misconduct incidents.

The model's reasoning: The paragraph recounts the paper's mortality finding but explicitly labels it incorrect and explains the sampling bias and subsequent retraction.

Where the model says the article warns about this paper:

• In the same section (Effects of mutations): In June 2019, researchers incorrectly suggested that the purportedly genetically edited humans may have been mutated in a way that shortens life expectancy.
• In the same section (Effects of mutations): However, the article overlooked sampling bias in UK Biobank's data, resulting in an erroneous interpretation, and was retracted four months later.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-21363 / 59162859 / #1, read from revision 1374845546 of He Jiankui affair.

The model's reasoning: The finding is reported as a claim of an explicitly retracted study, and the paragraph explains the sample contamination.

Where the model says the article warns about this paper:

• In the same section (Origin hypotheses): now-retracted study
• In the same section (Origin hypotheses): After a re-analysis because of doubts, the team retracted the article on 20 December 2022, due to contamination of the samples.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 0.99. Finding RW-41581 / 69363482 / #1, read from revision 1374223939 of SARS-CoV-2 Omicron variant.

The model's reasoning: The study's claimed finding is reported in a paragraph that explicitly identifies its data as falsified and the paper as retracted.

Where the model says the article warns about this paper:

• In the same section (Opposition): In 2011, Andrew Wakefield, a leading proponent of the theory that MMR vaccine causes autism, was found to have been financially motivated to falsify research data and was subsequently stripped of his medical license.
• In the same section (Vaccination and autism): The article was widely criticized for lack of scientific rigor and it was proven that Wakefield falsified data in the article.
• In the same section (Vaccination and autism): In 2004, 10 of the original 12 co-authors (not including Wakefield) published a retraction of the article and stated the following: "We wish to make it clear that in this paper no causal link was established between MMR vaccine and autism as the data were insufficient."
• In the same section (Vaccination and autism): In 2010, The Lancet officially retracted the article, stating that several elements of the article were incorrect, including falsified data and protocols.
• In the same section (Vaccination and autism): even though the article was shown to be fraudulent and was heavily retracted
• In the same section (Vaccination and autism): Even though there is much evidence to go against the Wakefield study and retractions were published by most of the co-authors, many people continue to believe and base decisions on the study as it still lingers in their memory.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-4036 / 32473 / #1, read from revision 1372277133 of Vaccination.

The model's reasoning: This sentence concerns the paper's misconduct rather than relying on its findings; the paragraph also explicitly describes its retraction.

Where the model says the article warns about this paper:

• In the same section (Opposition): In 2011, Andrew Wakefield, a leading proponent of the theory that MMR vaccine causes autism, was found to have been financially motivated to falsify research data and was subsequently stripped of his medical license.
• In the same section (Vaccination and autism): The article was widely criticized for lack of scientific rigor and it was proven that Wakefield falsified data in the article.
• In the same section (Vaccination and autism): In 2004, 10 of the original 12 co-authors (not including Wakefield) published a retraction of the article and stated the following: "We wish to make it clear that in this paper no causal link was established between MMR vaccine and autism as the data were insufficient."
• In the same section (Vaccination and autism): In 2010, The Lancet officially retracted the article, stating that several elements of the article were incorrect, including falsified data and protocols.
• In the same section (Vaccination and autism): even though the article was shown to be fraudulent and was heavily retracted
• In the same section (Vaccination and autism): Even though there is much evidence to go against the Wakefield study and retractions were published by most of the co-authors, many people continue to believe and base decisions on the study as it still lingers in their memory.

The reference is tagged {{Retracted}}, with intentional=yes.

The model's confidence: 1.00. Finding RW-4036 / 32473 / #2, read from revision 1372277133 of Vaccination.