Search bioRxiv⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/2025.08.29.673016

Revealing the Paper Mill Iceberg: AI-Based Screening of Cancer Research Publications.

Scancar, B.·Byrne, J. A.·Causeur, D.·Barnett, A. G.
View original source ↗Export metadata ↓

Abstract

ObjectivesTo train and validate a machine learning model to distinguish paper mill publications from genuine cancer research articles, and to screen the cancer research literature to assess the prevalence of papers that have textual similarities to paper mill papers. DesignMethodological study applying a BERT-based text classification model to article titles and abstracts. SettingRetracted paper mill publications listed in the Retraction Watch database were used for model training. The cancer research corpus was screened by the model, using the PubMed database restricted to original cancer research articles published between 1999 and 2024. ParticipantsThe model was trained on 2,202 retracted paper mill papers and validated on independent data collected by image integrity experts. A total of 2.6 million cancer research papers were screened. Main outcome measuresClassification performance of the model. Prevalence of papers flagged as similar to retracted paper mill publications with 95% confidence intervals and their distribution over time, by country, publisher, cancer type, research area, and within high-impact journals (Decile 1). ResultsThe model achieved an accuracy of 0.91. When applied to the cancer research literature, it flagged 9.87% (95% CI 9.83 to 9.90) of papers and revealed a large increase in flagged papers from 1999 to 2024, both across the entire corpus and in the top 10% of journals by impact factor. Over 170,000 papers affiliated with Chinese institutions were flagged, accounting for 35% of Chinese cancer research articles. Most publishers had published substantial numbers of flagged papers. Flagged papers were overrepresented in fundamental research and in gastric, bone, and liver cancer. ConclusionsPaper mills are a large and growing problem in the cancer literature and are not restricted to low impact journals. Collective awareness and action will be crucial to address the problem of paper mill publications.

Source connections

Explore related subjects

cancer biologyPreprint

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Scancar, B., Byrne, J. A., Causeur, D., Barnett, A. G.. 2025-09-03. Revealing the Paper Mill Iceberg: AI-Based Screening of Cancer Research Publications.. https://doi.org/10.1101/2025.08.29.673016

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Regulation of the Epstein-Barr Epigenome by Histone Variant MacroH2A

Histone variants play an important role in regulating chromatin dynamics of Epstein-Barr Virus (EBV). Here, we investigate the role of histone variant macroH2A in EBV chromatin structure, gene expression and copy number control during latent infection. ChIP-seq analysis in EBV+ SNU719 gastric carcinoma and Mutu I Burkitt lymphoma cells revealed that macroH2A was most enriched at repetitive DNA elements, including oriLyt, BHLF1, LF3 and terminal repeat (TR) regions of the EBV genome. MacroH2A depletion in EBV-associated gastric cancer cell line SNU719 led to a small increase in some viral lytic genes and increase in viral DNA copy number. RNAseq transcriptomics after macroH2A depletion identified major perturbations in cellular gene pathways, including beta-catenin/TCF signaling and chromatin remodeling. Depletion of macroH2A during EBV lytic reactivation reduced EBV lytic gene activation, but increased EBV DNA copy number. Previous study demonstrated a colocalization and cooperative binding between histone variant H2A.Z and EBNA1. We now show that depletion of macroH2A causes an increase in both H2A.Z and EBNA1 at oriP and cellular binding sites for EBNA1. We also found that gammaH2AX localizes to oriP and that this depends on macroH2A. Finally, we show that CRISPR/Cas9 knockout of macroH2A increases the replication and episome maintenance of oriP-containing plasmids. Taken together, these findings indicate that macroH2A localizes to the repetitive elements of EBV, including oriLyt, and can function to limit H2A.Z and EBNA1 binding at oriP, which alters DNA damage signaling at oriP and limits episomal replication. These findings provide a novel role for macroH2A in maintaining EBV copy number and genome integrity.

cancer biology↗
bioRxiv2026

Epigenomic circuits shaping cellular heterogeneity in pleural mesothelioma

Pleural mesothelioma (PM) is an aggressive and lethal malignancy with current treatments offering limited clinical benefit. In this study, we utilized single-cell chromatin accessibility and transcriptomic profiling of patient-matched samples to infer the epigenomic regulation underlying cellular heterogeneity in PM. High-resolution epigenomic profiling suggested transcriptional regulators with putative pro-tumoral roles in mesothelioma, whose function was experimentally tested using genetic perturbations across four PM cell lines. Furthermore, we observed several putative associations between epigenomic circuits and macrophage ontogeny, a sarcomatoid-like neoplastic state, and a fetal-like endothelial cell state. Our analysis also showed similarities between the chromatin accessibility landscapes of exhausted CD8 T and KLRC1 natural killer cells and identified NR4A2 as a potential shared regulator of dysfunction in both lymphocyte subsets. We further identified a state-specific distal enhancer near NR4A2 and experimentally validated its regulation of NR4A2 expression in Jurkat cells. Together, we provide an epigenomic map of PM and nominate candidate transcription factors and regulatory elements for future functional investigation, including their potential relevance to epigenetic editing strategies to limit tumor growth and enhance anti-tumor immunity.

cancer biology↗
bioRxiv2026

Multimodal Liquid Biopsy Improves Longitudinal Circulating Tumor DNA Detection in Children with Central Nervous System Tumors

Background: Circulating tumor DNA (ctDNA) is a promising biomarker for pediatric central nervous system (CNS) tumors, but limitations in sensitivity and universally accepted detection criteria remain. We evaluated multimodal ctDNA detection using a novel composite scoring framework integrating mutation, copy number, and fragmentomic analysis in longitudinal plasma and CSF samples from children with CNS tumors. Methods: Tumor, blood and/or CSF samples were collected from pediatric CNS tumor patients. DNA was extracted from plasma, CSF, tumor, and whole blood. All samples underwent targeted and whole genome sequencing. Plasma and CSF samples were evaluated for ctDNA positivity--defined by a composite score [≥]2--utilizing a tri-modal approach of whole genome sequencing for copy number alterations and fragmentomics and targeted sequencing. Results: The cohort included 28 patients: medulloblastoma (39%), diffuse midline glioma (25%), high-grade glioma (18%), pilocytic astrocytoma (11%), and ependymoma (7%). A total of 140 plasma and 27 CSF samples were collected longitudinally. Combining all three modalities identified ctDNA in 70% of CSF and 17% of plasma samples, exceeding rates achieved by any single modality alone. Post-diagnosis plasma ctDNA positivity was significantly associated with worse progression-free and overall survival. CSF demonstrated significantly higher ctDNA signal intensity than plasma across all assay modalities. Conclusions: A multimodal composite scoring approach improves longitudinal ctDNA detection in pediatric CNS tumors. Although CSF remains the most sensitive biofluid, plasma ctDNA reflects disease progression in select patients. These findings support integrated, multimodal liquid biopsy strategies for disease monitoring and highlight the need for prospective validation of standardized detection criteria.

cancer biology↗