Digital magic, or the dark arts of the 21 st century—how can journals and peer reviewers detect manuscripts and publications from paper mills?

The similarities between and errors within SGK articles could reflect the undeclared involvement of organisations such as paper mills.

“We proposed that the similarities between and errors within SGK articles could reflect the undeclared involvement of organisations such as paper mills (13), which have been alleged to mass-produce fraudulent manuscripts for publication (24, 25).”

Producing many gene research manuscripts at minimal cost could involve writers with either an incomplete understanding of the experiments that they are describing and/or limited time for quality control.

“We also proposed that producing many gene research manuscripts at minimal cost could involve writers with either an incomplete understanding of the experiments that they are describing and/or limited time for quality control (8, 24).”

The production of manuscripts at scale could underpin unusual degrees of similarity between publications, as well as features such as superficial explanations for the analysis of particular genes, and generic experimental approaches.

“We proposed that the production of manuscripts at scale could underpin unusual degrees of similarity between publications, as well as features such as superficial explanations for the analysis of particular genes, and generic experimental approaches (8, 13, 24, 25, 26).”

These conditions could lead to wrongly identified nucleotide sequences being a feature of gene research articles from paper mills, where incorrect sequences could also be reused across different manuscripts.

“These conditions could lead to wrongly identified nucleotide sequences being a feature of gene research articles from paper mills (8, 24) where incorrect sequences could also be reused across different manuscripts (11, 12, 13).”

Unusual reagent errors combined with implausible or impossible results could flag that some affected articles are fraudulent.

“Such unusual reagent errors combined with implausible or impossible results could flag that some affected articles are fraudulent (8, 11, 12, 13, 24, 26).”

Wrongly identified nucleotide sequences in the context of repetitive and superficially justified gene research could reflect the external involvement of organisations such as paper mills.

“We have previously proposed that wrongly identified nucleotide sequences in the context of repetitive and superficially justified gene research could reflect the external involvement of organisations such as paper mills (8, 13, 24, 26).”

The availability of many human genes for experimental analysis, combined with research drivers that favour the continued investigation of genes of known function, may unwittingly provide an extensive source of topics around which gene research articles can be fraudulently created.

“The availability of many human genes for experimental analysis, combined with research drivers that favour the continued investigation of genes of known function (57, 58, 59), may unwittingly provide an extensive source of topics around which gene research articles can be fraudulently created (8, 24).”

Clients of paper mills are likely to be academics, research students and clinicians who do not have the time, facilities and/or training to conduct the genuine research that is required of them.

“Clients of paper mills are likely to be academics, research students and clinicians who do not have the time, facilities and/or training to conduct the genuine research that is required of them (10,11,32,34,35).”

Paper mills are commercial entities that sell undeclared services in relation to research manuscripts and publications.

“'paper mill' is applied to commercial entities that (also) sell undeclared services in relation to research manuscripts and publications (8–12,32,33).”

Paper mills are alleged to operate in many countries and offer a variety of questionable and/or illegal services to a range of clients across different research disciplines.

“Paper mills are alleged to operate in many countries and to offer a variety of questionable and/or illegal services to a range of clients across different research disciplines (8–10,32,33).”

Clients are believed to be drawn to paper mills by imbalances between the pressures and opportunities to publish research in different settings.

“Clients are believed to be drawn to paper mills by imbalances between the pressures and opportunities to publish research in different settings (8,10,11,32,35,43).”

The current inability to achieve timely corrections of the published literature at scale specifically disadvantages genuine research, while providing an invaluable asset for paper mills.

“In summary, the current inability to achieve timely corrections of the published literature at scale specifically disadvantages genuine research, while providing an invaluable asset for paper mills (10).”

Paper mill employees are alleged to undertake manuscript submission, including simultaneous submission to multiple journals.

“Paper mill employees are also alleged to undertake manuscript submission, which can involve the simultaneous submission of manuscripts to multiple journals, presumably to increase the likelihood and speed of manuscript acceptance (10,32).”

Paper mill manuscripts that describe laboratory research will frequently include falsified or fabricated experimental results.

“It is predicted that paper mill manuscripts that describe laboratory research will frequently include falsified or fabricated experimental results, due to genuine laboratory research being difficult, expensive and slow (10,27,32,36,38).”

Targeted human gene research provides several scaling factors that could enable the production of many individual manuscripts.

“Targeted human gene research provides several scaling factors that could enable the production of many individual manuscripts (10,27,56) (Figure 4).”

The value of accelerating this rate-limiting step has undoubtedly been recognized by paper mills, through strategies such as editorial and peer review manipulation, and submissions of the same or very similar manuscripts to multiple journals.

“The value of accelerating this rate-limiting step has undoubtedly been recognized by paper mills, through strategies such as editorial and peer review manipulation, and submissions of the same or very similar manuscripts to multiple journals (8,10,32,33,35,63,73).”

Publishers and individual journals are becoming increasingly aware of paper mills, but this awareness is unlikely to extend to researchers, who generally expect low rates of outright research fraud.

“While publishers and individual journals are becoming increasingly aware of paper mills (8,10,32,36,62–73), this awareness is unlikely to extend to researchers, who generally expect low rates of outright research fraud (7,8).”

One repeatedly noted feature of suspected paper mill manuscripts and articles is the absence of mechanistic hypotheses to link genes, systems, and experimental approaches through novelty-based research justifications.

“One repeatedly noted feature of suspected paper mill manuscripts and articles is the absence of mechanistic hypotheses to link gene(s), systems, and experimental approaches through novelty-based research justifications (10,65,71).”

Journal requirements that insist on focussed research justifications and technical approaches could provide some level of deterrence, at least in the short-term.

“journal requirements that insist on focussed research justifications and technical approaches could provide some level of deterrence (10,71,89–92), at least in the short-term.”

Institutionally-supported education that provides information about features of problematic research and paper mills will allow trainees and researchers to more critically assess the literature that they use to inform their projects and experiments.

“Institutionally-supported education that provides information about features of problematic research and paper mills (Table 2) will allow trainees and researchers to more critically assess the literature that they use to inform their projects and experiments (8,10,27,32).”

Identification of verifiable errors such as wrongly identified nucleotide sequences or misidentified or contaminated cell line models could immediately precipitate the publication of an indexed editorial note or expression of concern.

“The identification of verifiable errors such as wrongly identified nucleotide sequences (38,39,54,56) or misidentified or contaminated cell line models (97) could immediately precipitate the publication of an indexed editorial note (10) or expression of concern.”

Widespread fraudulent gene research will encourage the financial support of unproductive research directions.

“Widespread fraudulent gene research will encourage the financial support of unproductive research directions”

Requirements to preprint manuscripts where both the submission date and journal are clearly displayed could reduce the burden of duplicate submissions on journals and peer reviewers.

“Requirements to preprint manuscripts where both the submission date and journal are clearly displayed (Table 3) could reduce the burden of duplicate submissions on journals and peer reviewers (10,72)”

At the time of manuscript submission, paper mills can increase manuscript acceptance rates by simply submitting the same manuscripts to multiple journals.

“At the time of manuscript submission, paper mills can increase manuscript acceptance rates by simply submitting the same manuscripts to multiple journals”

Paper mills will also be selectively disadvantaged by geometrically increasing the rate and scale of post-publication corrections.

“Paper mills will also be selectively disadvantaged by geometrically increasing the rate and scale of post-publication corrections (10) (Table 3).”

Submitting the same manuscripts to multiple journals is highly wasteful of editorial and peer review resources.

“which is highly wasteful of editorial and peer review resources (10,32,36,72).”

To date, manuscripts or articles that may have been produced with assistance from paper mills have been recognised by researchers analysing unexpected literature trends and/or verifiable errors, and by journal editors and experts describing unusual manuscript or publication features.

“To date, manuscripts or articles that may have been produced with assistance from paper mills has been recognised by researchers analysing unexpected literature trends (38,85,86) and/or verifiable errors (38,39,56) and by journal editors and experts describing unusual manuscript or publication features (10,36,37,62–73,82).”

Requirements to preprint manuscripts where both the submission date and journal are clearly displayed could reduce the numbers of paper mill manuscripts that are accepted for publication.

“and the numbers of paper mill manuscripts that are accepted for publication.”

Widespread fraudulent gene research will reduce confidence in research and the scientific method.

“and reduce confidence in research and the scientific method (10,33,56).”

Widespread fraudulent gene research will slow research translation through opportunity costs.

“slow research translation through opportunity costs”

Byrne and Christopher (2020) have identified two types of fraudulent images: invented images and stock images.

“Byrne and Christopher (Byrne and Christopher, 2020) have identified two types of fraudulent images: invented images and stock images.”

Fabricated research papers are readily available for purchase by scientists driven by the need to publish more papers to further their careers.

“Moreover, it appears that fabricated research papers are readily available for purchase by scientists driven by the need to publish more papers to further their careers (see Byrne and Christopher, 2020).”

Stock images tend only to be spotted by eagle-eyed readers if repeated in multiple submissions or publications.

“They tend only to be spotted by eagle-eyed readers if repeated in multiple submissions or publications.”

Invented images are typically western blots, and might include placing individual bands onto false backgrounds, non-linear adjustments, the splicing of multiple images to suggest they come from the same gel, duplicated bands and lanes, and grouping or consolidation of the data (e.g. removal of lanes).

“Invented images are typically western blots, and might include placing individual bands onto false backgrounds, non-linear adjustments, the splicing of multiple images to suggest they come from the same gel, duplicated bands and lanes, and grouping or consolidation of the data (e.g. removal of lanes) (see Box 1).”

Stock images are more difficult to detect and show none of the manipulation likely from the production of invented images.

“Stock images are more difficult to detect and show none of the manipulation likely from the production of invented images.”