Dark Satanic Papermills #1

This post was earlier cross-posted at Leonid Schneider's site, hence the unfrivolous tone. The version there is improved by Leonid's editing, background details and frame-story. Also a breakdown of papers by journal, and Tiger's map. Dr Bik has written her own more-coherent blogpost with other angles on the story. Drs Byrne and Christopher subsequently published an independent examination of the papermill phenomenon, including some of the same cases.

News outlets picked up on the story; publishers and journals are investigating. The Chinese government announced a new policy of not rewarding academics with a bounty for papermilled publications in English-language journals. Also, Paris, Rome, Munich and New York are burning.

Here are some frames from a computer-animated version of Star Wars IV, plotted with an old line-printer (that's what we did for entertainment in the late 1970s). The Death Star down in each frame's quadrant Q3 is shooting out bolts of planet-smashing energy to the right through Quadrant Q4, while disciplined flotillas of Rebel Alliance X-Wing starfighters are swooping down through Q1 and Q2.

No, I made that up. These hairballs are purportedly FACS scatterplots where each dot is a cell, located by its surface-protein measurements: Figures from Tang et al. (2020) [11]. Surprisingly many cells yielded identical values in these independent experiments!

Other papers have reported further frames from the animation, overlapping in the same way. Fig 2 from Ren, Xu & Xu (2019) [6]:

Figures 2E and 2F from Xiao and Tian (2019) [8]:

What to make of this phenomenon? Perhaps these Western Blots from the same papers will assist in the search for an explanation.



When blobs of protein are formed on an electrophoresis gel, blotted onto a membrane and then picked out with antibodies, they can take many forms and shapes, depending on the distortions as they were force-marched across the gel... depending also on their shape when they were streaked out onto the starting-line. They can bunch up as they travel, or spread out sideways into neighbouring lanes, so the spaces between blobs varies with molecular weight (i.e. travel distance). They should display the same general profile when modified, superimposed forms of a single protein are resolved using specific antibodies. They should not cavort like Zany Portuguese Sardines.

Must credit Bert Fegg's Nasty Book

Thus skepticism is understandable when one encounters Figures where the lane spacing is constant and bands of all molecular weights are random but interchangeable morphological mixtures. Especially when the background texture is identical when emphasised by increasing the contrast; not only between bands but between Figures, and even between papers... manuscripts from non-overlapping teams of researchers, at unconnected hospitals and research centres from far-flung provinces in China.

Indigofera Tanganyikensis reported the first two instances of this phenomenon (Liu et al, 2019 [13]; Liu et al., 2020 [14]). Then Clone Ranger Elisabeth Bik (a.k.a. Obik-wan Kenobik) joined the chase. Currently they have flagged over 30 430 of these sardine scrapbooks in PubPeer threads, too many to list here so TigerBB8 and I scraped off the details and arranged them into a spreadsheet. Most author-names are unique although a few research teams have been back for a second bite of the cherry.

In theory the corresponding authors of each paper have been notified and invited to join the discussion of their work (an automagic feature of the PubPeer software), though no-one has shown up yet. In some cases the invitation may not have reached the authors, due to their choice of Gmail e-addresses, which are no longer accessible in mainland China.

BrendaWillingham12192@gmail.com (Huang & Jin, 2018 [1])

nancycook4925@gmail.com (Ma et al., 2018 [2]) 

marildivisio25318@gmail.com (Chen et al., 2018 [4])

RosettajKirkland3814@gmail.com (Zhang et al. 2018 [5])

CaseyPeiffer8311@gmail.com (Zhang & Liu, 2019 [10])

Nancy Cook's Fig 7 from [2] has the same band background as Brenda Willingham's Fig 5C from [1]

Now I am in no position to criticise anyone for adopting a playful pseudonym for correspondence, but these are not as academic as I would expect. The sources of the nyms are not obvious, apart from Casey Peiffer, who is a 9/11 Truther of some renown. At any rate the journal has evidently managed to stay in contact with whoever submitted them. This is an improvement on another early example in this tradition, Liu et al. (2017) [3], which was retracted when the editors could not contact the authors for the final authorisation. Not because of this:

The topics of these studies tend to conform to a template, helpfully summarised by Dr Bik in Mad-Libs format.

Phytochemicals from the materia medica of Traditional Chinese Medicine are common -- Astrogaloside, Baicalein, Salidroside. I begin to wonder if there are any TCM-approved phytochemicals which don't inhibit the proliferation and migration of medulloblastoma Daoy cells, in the hands of staff from Jining #1 People's Hospital.

The majority of papers appeared in the Journal of Cellular Biochemistry from the Wiley stable, though not exclusively so (Biomedicine & Pharmacotherapy features as well, and Artificial Cells Nanomedicine and Biotechnology, and a few others). This may change. I would not be surprised if the PubPeer commentariat simply focussed on recent issues of J.Cell.Biochem. first, so we should not succumb to Selection Bias.

J.Cell.Biochem. has become popular among China-based molecular-biology researchers as an outlet for their work, and has a respectable Citations Index of 3.448 (though how much of that citational credit flows in from other journals remains to be seen). Despite this, my initial hope that the editors would consult experts to review submissions has been dispelled by their blithe acceptance of the egregious images above. Perhaps the peer-reviewers and contributors are collectively engaged in a Post-Modern project of redefining what Western Blots and FACS plots should look like, forging new expectations and conventions.

Two papers featured what purport to be Transwell Invasion Assays, while more closely resembling exercises in fumage and collage; the reviewers also thought these were fine (Zhang et al, 2019 [7]; Zhuang, Liu & Wu, 2019 [9]). Despite non-overlapping author lists, the images of both papers were built from a single small repertoire of fumage motifs.

Notably, both papers included more frames from the Death Star animation, which is why they are included here despite the absence of zany-sardine Western Blots.

Elsewhere the reviewers signed off on this Declaration of Data Availability, puzzling in its uncertainty: "The data set(s) supporting the conclusions of this article is(are) included within the article" (Liu et al., 2020 [12]).

The article contained no raw data, lacking even the number of repetitions behind the graphs, though the statement could still be true if in fact the data sets don't exist. Or if they exist but contradict the conclusions.

Setting the journal aside, the sources of these papers are not geographically diverse. The current list shows a concentration in Shandong Province, in Jining and Qingdao and a few other cities. Changchun, in Jilin Province, is a smaller hotspot. Those two groups trying TCM on medulloblastoma at the Department of Paediatrics, Jining No. 1 People's Hospital seem to be working independently; they should get in touch and coordinate their efforts. [4], a.k.a. marildivisio25318, I'd like you to meet [5], a.k.a. RosettajKirkland3814... but maybe you have already met.


All this has a parsimonious explanation in a policy in China's medical sector, whereby promotion for everyone -- clinicians and researchers alike -- depends on publishing academic papers. Of course the average overworked doctor has neither the time, the training nor the resources to apply transfection technology to the interaction between triptolide, microRNA-1462 and JNK pathways in osteosarcoma, then write it up for publication, so they pay someone else to make it all up. And if you were familiar with the conventions of the academic genre, and had the software for plotting made-up statistics, wouldn't you take advantage of this market opportunity by setting up a papermill?

Presumably the persons behind this service are relying on word-of-mouth to advertise their service, hence the geographical concentration. They have friends or members outside the mainland, allowing them to access the bogus Gmail accounts and keep in touch with the journal. One can speculate that there is someone within the editorial structure of J.Cell.Biochem. who is in on the deal and has suborned the peer-review process to ensure that manuscripts are only sent to friendly or fictitious reviewers (the same deduction might also follow for other journals; we await further developments).

Are there any victims in this curious market-driven ecology, where the nominal authors get the CV-stuffers they need and the actual authors get paid? Opinions are divided. There is the danger that genuine biomedical researchers in China might be disadvantaged if they foolishly published in J.Cell.Biochem., only to have their work dismissed and ignored. It would be helpful if the journal introduced some sort of flag to let readers know which of their articles are based on real experiments and which ones are information-free fabrications, forged to meet contractual / promotional obligation.

SOURCES:

[1]. "ZIC2 promotes viability and invasion of human osteosarcoma cells by suppressing SHIP2 expression and activating PI3K/AKT pathways", Shuaihao Huang, Anmin Jin.
Journal of Cellular Biochemistry (2018), doi: 10.1002/jcb.26387 [PubPeer].

[2]. "MicroRNA-29a inhibits proliferation and motility of schwannoma cells by targeting CDK6", Ji Ma, Tengfei Li, Huifeng Yuan, Xinwei Han, Shaofeng Shui, Dong Guo, Lei Yan.
Journal of Cellular Biochemistry (2018) doi: 10.1002/jcb.26426 [PubPeer].

[3]. "Retraction : HCFU inhibits cervical cancer cells growth and metastasis by inactivating Wnt/β‐catenin pathway", Ping Liu, Shuying Ma, Hua Liu, Huazhen Han, Shanshan Wang,
Journal of Cellular Biochemistry (2017) doi: 10.1002/jcb.26570 [PubPeer].

[4]. "Ginsenoside Rh2 inhibits proliferation and migration of medulloblastoma Daoy by down-regulation of microRNA-31", Yan Chen, Hong Shang, Shunli Zhang, Xiaohong Zhang.
Journal of Cellular Biochemistry (2018) doi: 10.1002/jcb.26716 [PubPeer].

[5]. "Triptolide inhibits the proliferation and migration of medulloblastoma Daoy cells by upregulation of microRNA-138", Haifang Zhang, Hui Li, Zhenguo Liu, Ang Ge, Enyu Guo, Shuxia Liu, Zhiping Chen.
Journal of Cellular Biochemistry (2018) doi: 10.1002/jcb.27307 [PubPeer].

[6]. "Salidroside represses proliferation, migration and invasion of human lung cancer cells through AKT and MEK/ERK signal pathway", Mei Ren, Wenjing Xu, Tao Xu
Artificial Cells Nanomedicine & Biotechnology (2019) doi: 10.1080/21691401.2019.1584566 [PubPeer].

[7]. "Downregulation of microRNA-1469 promotes the development of breast cancer via targeting HOXA1 and activating PTEN/PI3K/AKT and Wnt/β-catenin pathways", Yonghui Zhang, Jing Fang, Hongmeng Zhao, Yue Yu, Xuchen Cao, Bin Zhang.
Journal of Cellular Biochemistry (2019) doi: 10.1002/jcb.27786 [PubPeer].

[8]. "Knockdown of long noncoding RNA HOTAIR inhibits cell growth of human lymphoma cells by upregulation of miR‐148b", Xianxian Zhao, Xiaoyan Tian
Journal of Cellular Biochemistry (2019) doi: 10.1002/jcb.28500 [PubPeer].

[9]. "Upregulation of long noncoding RNA TUG1 contributes to the development of laryngocarcinoma by targeting miR‐145‐5p/ROCK1 axis", Shenfa Zhuang, Fengxian Liu, Pingping Wu.
Journal of Cellular Biochemistry (2019) doi: 10.1002/jcb.28614 [PubPeer].

[10]. "Sapylin inhibits lung cancer cell proliferation and promotes apoptosis by attenuating PI3K/AKT signaling", Lin Zhang, Benhong Liu
Journal of Cellular Biochemistry (2019) doi: 10.1002/jcb.28729 [PubPeer].

[11]. "Long noncoding RNA MEG3 deteriorates inflammatory damage by downregulating microRNA‐101a", Shouyi Tang, Junxia Han, Hui Jiao, Jingna Si, Yingying Liu, Jinlong Wang
Journal of Cellular Biochemistry (2020) doi: 10.1002/jcb.29415 [PubPeer].

[12]. "Silence of cZNF292 suppresses the growth, migration, and invasion of human esophageal cancer Eca‐109 cells via upregulating miR‐206", Zengjia Liu, Guiju Hu, Yan Zhao, Zuorun Xiao, Mingzhe Yan, Mei Ren.
Journal of Cellular Biochemistry (2020) doi: 10.1002/jcb.29458 [PubPeer].

[13]. "Circular RNA ACR relieves high glucose‐aroused RSC96 cell apoptosis and autophagy via declining microRNA‐145‐3p", Ying Liu, Xiaoqing Chen, Jingjing Yao, Jing Kang
Journal of Cellular Biochemistry (2019) doi: 10.1002/jcb.29568 [PubPeer].

[14]. "Intermedin attenuates cardiomyocytes hypoxia‐injury through upregulating long noncoding RNA MALAT1", Long Liu, Haiming Xu, Jingze Zhang, Liquan Yin, Qini Zhao
Journal of Cellular Biochemistry (2020) doi: 10.1002/jcb.29642 [PubPeer].