Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 87 of the 93 references it cites were found in a registry or an index; 90 of the 150 sources its claims cite were checked for support.

Identification of human gene research articles with wrongly identified nucleotide sequences

Open access

Published version of a preprint

Abstract

Nucleotide sequence reagents underpin molecular techniques that have been applied across hundreds of thousands of publications. We have previously reported wrongly identified nucleotide sequence reagents in human research publications and described a semi-automated screening tool Seek & Blastn to fact-check their claimed status. We applied Seek & Blastn to screen >11,700 publications across five literature corpora, including all original publications in Gene from 2007 to 2018 and all original open-access publications in Oncology Reports from 2014 to 2018. After manually checking Seek & Blastn outputs for >3,400 human research articles, we identified 712 articles across 78 journals that described at least one wrongly identified nucleotide sequence. Verifying the claimed identities of >13,700 sequences highlighted 1,535 wrongly identified sequences, most of which were claimed targeting reagents for the analysis of 365 human protein-coding genes and 120 non-coding RNAs. The 712 problematic articles have received >17,000 citations, including citations by human clinical trials. Given our estimate that approximately one-quarter of problematic articles may misinform the future development of human therapies, urgent measures are required to address unreliable gene research articles.

Overview

CoChat Summary

TL;DR

Using Seek & Blastn with manual verification, this study screened >11,700 human gene research articles across five corpora and identified 712 problematic articles containing 1,535 wrongly identified nucleotide sequences. These articles span 78 journals, have accumulated >17,000 citations, and almost all remain uncorrected.

Key findings

  • S&B screened >11,700 articles, flagged 3,423 for manual verification, and confirmed 712 problematic articles with 1,535 wrongly identified sequences across 31 publishers and 78 journals.
  • Problematic article rates were 75/174 (43%) in the SGK corpus, 31/50 (62%) in the miR-145 corpus, 51/100 (51%) in the cisplatin + gemcitabine corpus, 128/742 flagged Gene articles (17%), and 436/1,709 flagged Oncology Reports articles (26%).
  • Most incorrect sequences were (RT-)PCR reagents (1,301/1,535, 85%), followed by si/shRNAs (226/1,535, 15%); most problematic articles in targeted and Oncology Reports corpora were from China, often hospital-affiliated.
  • The 712 problematic articles had been cited 17,183 times by March 2021; only 2% (11/712) were retracted, three had expressions of concern, and average approximate potential to translate ranged from 15% to 35% across corpora.

Methods

  • Semi-automated Seek & Blastn (S&B) extracted nucleotide sequences and text descriptors, used Blastn to predict targeting/non-targeting status, flagged contradictions, and manually verified flagged outputs; >13,700 nucleotide sequences were checked.
  • Five corpora were screened: three targeted corpora (SGK articles for 17 human genes; miR-145 articles; cisplatin and gemcitabine articles) and two journal corpora (Gene 2007–2018; open-access Oncology Reports 2014–2018).
  • Analyses included country and institutional affiliation, post-publication notices, citation counts, gene knowledge-base indexing, gene2pubmed linkage, Fisher’s exact tests, Mann–Whitney U tests, and approximate potential to translate (APT).

Limitations

  • Stated: not all articles in Gene and Oncology Reports were manually verified (10% and 45%, respectively), so some problematic articles may have been missed; Blastn’s low but measurable false-negative rate means some flagged non-targeting reagents could be false positives, though multiple errors per article provide some protection.
  • Stated: targeted corpora were derived from search terms and index articles previously associated with problematic articles, mostly from China, so rates may not reflect the wider literature; miR-145 and C+G corpora were not exhaustive, and journal selection may have increased detection of problematic articles from China and hospital-affiliated authors.
  • Evident from the text: the study is observational and cannot prove fraud or paper-mill involvement; country and institution associations are descriptive, and reagent errors were not validated by wet-lab experiments.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.26508/lsa.202101203

Cites

87 of 87 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1. Diagram describing the five literature corpora screened by S&B. For each corpus (top row), the diagram shows the numbers of articles that were (i) screened by S&B (white), (ii) flagged by S&B with sequences manually verified (grey), and (iii) found to be problematic by describing at least one wrongly identified nucleotide sequence (dark grey). Total numbers of problematic articles and wrongly identified sequences are indicated below the diagram, corrected for duplicate articles between the corpora.
Figure 1. Diagram describing the five literature corpora screened by S&B. For each corpus (top row), the diagram shows the numbers of articles that were (i) screened by S&B (white), (ii) flagged by S&B with sequences manually verified (grey), and (iii) found to be problematic by describing at least one wrongly identified nucleotide sequence (dark grey). Total numbers of problematic articles and wrongly identified sequences are indicated below the diagram, corrected for duplicate articles between the corpora.CC BY · PubMed Central
Figure 2. Percentages of sequence identity error types in each corpus. Percentages of wrongly identified nucleotide sequence reagents that correspond to the three identity error types (y-axis) in each corpus (x-axis). Percentages corresponding to each error type are indicated, rounded to the nearest single digit. The numbers of incorrect sequences in each corpus are shown below the x-axis.
Figure 2. Percentages of sequence identity error types in each corpus. Percentages of wrongly identified nucleotide sequence reagents that correspond to the three identity error types (y-axis) in each corpus (x-axis). Percentages corresponding to each error type are indicated, rounded to the nearest single digit. The numbers of incorrect sequences in each corpus are shown below the x-axis.CC BY · PubMed Central
Figure 3. Percentages of wrongly identified nucleotide sequences that were either unique or repeated within each corpus. Percentages of wrongly identified sequences that were identified at least twice in any single corpus (black) are shown above each image, rounded to the nearest single digit. All other wrongly identified sequences were unique in the indicated corpus (grey). Numbers of wrongly identified sequences identified in each corpus are shown below each image.
Figure 3. Percentages of wrongly identified nucleotide sequences that were either unique or repeated within each corpus. Percentages of wrongly identified sequences that were identified at least twice in any single corpus (black) are shown above each image, rounded to the nearest single digit. All other wrongly identified sequences were unique in the indicated corpus (grey). Numbers of wrongly identified sequences identified in each corpus are shown below each image.CC BY · PubMed Central
Figure S1. Problematic Gene articles per year, according to country of origin and institutional affiliation type. Total numbers of problematic articles for each country/group of countries are shown in the upper left corner of each panel. (A) Numbers of problematic Gene articles (y-axes) per publication year (x-axes) according to country of origin, shown above each graph. Countries are shown in alphabetical order, from top left. (B) Numbers of problematic Gene articles (y-axis) per publication year (x-axis) from China (right panel) or all other countries (left panel). Articles affiliated with hospitals or other institution types are shown in blue or grey, respectively. Numbers of problematic articles per year are shown below each stacked bar graph.
Figure S1. Problematic Gene articles per year, according to country of origin and institutional affiliation type. Total numbers of problematic articles for each country/group of countries are shown in the upper left corner of each panel. (A) Numbers of problematic Gene articles (y-axes) per publication year (x-axes) according to country of origin, shown above each graph. Countries are shown in alphabetical order, from top left. (B) Numbers of problematic Gene articles (y-axis) per publication year (x-axis) from China (right panel) or all other countries (left panel). Articles affiliated with hospitals or other institution types are shown in blue or grey, respectively. Numbers of problematic articles per year are shown below each stacked bar graph.CC BY · PubMed Central
Figure 4. Percentages of problematic Gene and Oncology Reports articles according to hospital affiliation status and country of origin. Percentages of problematic Gene and Oncology Reports articles according to hospital affiliation status (y-axis) from either China or all other countries (x-axis). The journal and relevant date ranges of problematic articles are shown above each panel. Problematic articles that were (not) affiliated with hospitals are shown in blue (grey), respectively. Percentages shown have been rounded to the nearest single digit. Numbers of problematic articles from China or all other countries are indicated below the x-axis. For the comparisons shown in each panel, significantly higher proportions of problematic articles from China were affiliated with hospitals versus problematic articles from other countries (Fisher’s Exact test, P < 0.001).
Figure 4. Percentages of problematic Gene and Oncology Reports articles according to hospital affiliation status and country of origin. Percentages of problematic Gene and Oncology Reports articles according to hospital affiliation status (y-axis) from either China or all other countries (x-axis). The journal and relevant date ranges of problematic articles are shown above each panel. Problematic articles that were (not) affiliated with hospitals are shown in blue (grey), respectively. Percentages shown have been rounded to the nearest single digit. Numbers of problematic articles from China or all other countries are indicated below the x-axis. For the comparisons shown in each panel, significantly higher proportions of problematic articles from China were affiliated with hospitals versus problematic articles from other countries (Fisher’s Exact test, P < 0.001).CC BY · PubMed Central
Figure S2. Problematic Oncology Reports articles per year, according to country of origin and institutional affiliation type. Total numbers of problematic articles for each country/group of countries are shown in the upper left corner of each panel. (A) Numbers of problematic Oncology Reports articles (y-axes) per publication year (x-axes) according to country of origin, shown above each graph. Countries are shown in alphabetical order, from top left. (B) Numbers of problematic Oncology Reports articles (y-axis) per publication year (x-axis) from China (right panel) or all other countries (left panel). Articles affiliated with hospitals or other institution types are shown in blue or grey, respectively. Numbers of problematic articles per year are shown below each stacked bar graph.
Figure S2. Problematic Oncology Reports articles per year, according to country of origin and institutional affiliation type. Total numbers of problematic articles for each country/group of countries are shown in the upper left corner of each panel. (A) Numbers of problematic Oncology Reports articles (y-axes) per publication year (x-axes) according to country of origin, shown above each graph. Countries are shown in alphabetical order, from top left. (B) Numbers of problematic Oncology Reports articles (y-axis) per publication year (x-axis) from China (right panel) or all other countries (left panel). Articles affiliated with hospitals or other institution types are shown in blue or grey, respectively. Numbers of problematic articles per year are shown below each stacked bar graph.CC BY · PubMed Central
Figure 5. Percentages and numbers of problematic Gene and Oncology Reports articles per year. Percentages of all Gene or Oncology Reports articles that were found to be problematic (y-axis) per publication year (x-axis). The journal and relevant publication year ranges are shown above each panel. Problematic articles from China or all other countries are shown in orange or grey, respectively. Percentages shown are rounded to one decimal place. Total numbers of problematic articles per year are shown below each graph.
Figure 5. Percentages and numbers of problematic Gene and Oncology Reports articles per year. Percentages of all Gene or Oncology Reports articles that were found to be problematic (y-axis) per publication year (x-axis). The journal and relevant publication year ranges are shown above each panel. Problematic articles from China or all other countries are shown in orange or grey, respectively. Percentages shown are rounded to one decimal place. Total numbers of problematic articles per year are shown below each graph.CC BY · PubMed Central
Figure 6. Summary of (RT-)PCR primer pairings that involved at least one wrongly identified primer. For n = 851 primer pairs that were claimed to target particular genes/sequences (gene X) (left panel), one or both primers were predicted to be incorrect (right panel), either by targeting unrelated genes or sequences (gene Y or gene Z), or by having no predicted human target (no target). Numbers of primer pairs and affected articles are indicated below each incorrect primer pair category. Some problematic articles described more than one (category of) incorrect primer pairing. Left- or right-hand primers are not intended to indicate forward or reverse primer orientations.
Figure 6. Summary of (RT-)PCR primer pairings that involved at least one wrongly identified primer. For n = 851 primer pairs that were claimed to target particular genes/sequences (gene X) (left panel), one or both primers were predicted to be incorrect (right panel), either by targeting unrelated genes or sequences (gene Y or gene Z), or by having no predicted human target (no target). Numbers of primer pairs and affected articles are indicated below each incorrect primer pair category. Some problematic articles described more than one (category of) incorrect primer pairing. Left- or right-hand primers are not intended to indicate forward or reverse primer orientations.CC BY · PubMed Central
Figure 7. Numbers of past research articles that have studied human protein-coding genes in problematic articles. (A, B) Numbers (log base 10) of problematic articles (y-axis) versus past research articles (x-axis) for (A) primary protein-coding genes in problematic articles and (B) claimed protein-coding gene targets of wrongly identified reagents. Vertical dashed lines indicate the median number of research articles for protein-coding genes, with the associated interquartile range shown in grey. Subsets of protein-coding genes are highlighted in each panel.
Figure 7. Numbers of past research articles that have studied human protein-coding genes in problematic articles. (A, B) Numbers (log base 10) of problematic articles (y-axis) versus past research articles (x-axis) for (A) primary protein-coding genes in problematic articles and (B) claimed protein-coding gene targets of wrongly identified reagents. Vertical dashed lines indicate the median number of research articles for protein-coding genes, with the associated interquartile range shown in grey. Subsets of protein-coding genes are highlighted in each panel.CC BY · PubMed Central
Figure S3. Human protein-coding genes in problematic articles appear frequently in PubMed. (A, B) Numbers (log base 10) of PubMed articles (y-axis) for primary protein-coding genes in problematic articles (A) (green) or claimed protein-coding gene targets of wrongly identified reagents (B) (green), versus all other human protein-coding genes (A, B) (grey). Horizontal lines within box plots indicate median values, with box plots showing percentiles according to letter proportions, that is, ±25% percentile, ±(25 + 25/2)% percentile, ±(25 + 25/2 + 25/4)% percentile.
Figure S3. Human protein-coding genes in problematic articles appear frequently in PubMed. (A, B) Numbers (log base 10) of PubMed articles (y-axis) for primary protein-coding genes in problematic articles (A) (green) or claimed protein-coding gene targets of wrongly identified reagents (B) (green), versus all other human protein-coding genes (A, B) (grey). Horizontal lines within box plots indicate median values, with box plots showing percentiles according to letter proportions, that is, ±25% percentile, ±(25 + 25/2)% percentile, ±(25 + 25/2 + 25/4)% percentile.CC BY · PubMed Central
Figure 8. Clinical trial citations and approximate potential to translate (APT) for problematic articles. (A) Percentages of problematic articles that are cited at least once according to the NIH Open Citation Collection (y-axis), according to publication corpus (x-axis). Error bars indicate 95% confidence intervals of bootstrapped estimates of percentages. Numbers of problematic articles with at least one clinical citation are shown below the x-axis for each corpus. (B) Average APT for problematic articles (y-axis) according to publication corpus (x-axis). Error bars indicate bootstrapped 95% confidence intervals. Numbers of problematic articles for which the APT computed by iCite (88) are shown below the x-axis for each corpus.
Figure 8. Clinical trial citations and approximate potential to translate (APT) for problematic articles. (A) Percentages of problematic articles that are cited at least once according to the NIH Open Citation Collection (y-axis), according to publication corpus (x-axis). Error bars indicate 95% confidence intervals of bootstrapped estimates of percentages. Numbers of problematic articles with at least one clinical citation are shown below the x-axis for each corpus. (B) Average APT for problematic articles (y-axis) according to publication corpus (x-axis). Error bars indicate bootstrapped 95% confidence intervals. Numbers of problematic articles for which the APT computed by iCite (88) are shown below the x-axis for each corpus.CC BY · PubMed Central

Verification

Claims the paper makes

42 Valid · 2 Not valid · 6 Not checked · 174 Own results and interpretations

134 more claims · 19 with their cited sources checked

Create a free account to see all 224 claims and to verify the 1 reference still open.

See all claims — free account
  • Probably not a substantial claim. 39
  • It may state two separate claims as one. 57

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

16 more claims about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

5 of 8 · the rest aren’t loaded

3 more findings

Create a free account to see every finding of this check.

See all findings — free account

Cited works that were retracted

Works this paper cites that were later retracted, from the retraction notices publishers deposit. Major when the notice came out before this paper was published.

1 of 1

Citations

Cites

1 retracted

Show all 93 references

Research works printed but not linked: 2

  1. Clarivate Analytics (2019) Journal citation reports. https://jcr.clarivate.com/
  2. Hu, 2013, An empirical analysis on number and monetary value of ghostwritten papers in China, Curr Sci, 105, 1230

Cited by

See all 37

Metadata

Authors
Park, Y., West, R. A., Pathmendra, P., Favier, B., Stoeger, T., Capes-Davis, A., Cabanac, G., Labbé, C., & Byrne, J. A.
Affiliations
  • The University of Sydney (AU)
  • Translational Innovation in Medicine and Complexity (FR)
  • Northwestern University (US)
  • Faculty of Medicine and Health, The University of Sydney, Sydney, Australia
  • Centre National de la Recherche Scientifique (FR)
  • Institut polytechnique de Grenoble (FR)
Corresponding author
Byrne, Jennifer A
Venue
Life Science Alliance
Published
2022-01-12
Volume
5
Issue
4
Pages
e202101203
ISSN
2575-1077 (ISSN-L)
OpenAlex
W4206098416
Cited by
37
References
93
Citations by year
2026: 4 · 2025: 7 · 2024: 14 · 2023: 3 · 2022: 7
FWCI
6.49
Type
journal article
Language
en
Open access
Locations
  • Life Science Alliance (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • HAL (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • US Office of Research Integrity (ORIIR180038-01-00)
  • National Health and Medical Research Council of Australia (APP1184263)
  • National Science Foundation (1956338)
  • National Institutes on Aging (K99AG068544)
  • National Institute of Allergy and Infectious Diseases (AI135964)
Topics
  • Social Sciences › Library and Information Sciences › Academic Publishing and Open Access
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
  • Social Sciences › Education › Academic integrity and plagiarism
Keywords
human protein-coding genes, non-coding RNA, research integrity, citation impact, literature screening
MeSH terms
Scientific Experimental Error, Human Genetics, Base Sequence, Humans, Proteins, Publications, Genome, Human, Genetic Research

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-01, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-01 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Identification of human gene research articles with wrongly identified nucleotide sequences. https://app.cochat.ai/papers/10.26508/lsa.202101203
The paper
Park, Y., West, R. A., Pathmendra, P., Favier, B., Stoeger, T., Capes-Davis, A., Cabanac, G., Labbé, C., & Byrne, J. A. (2022). Identification of human gene research articles with wrongly identified nucleotide sequences. Life Science Alliance. https://doi.org/10.26508/lsa.202101203

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.