The brave women and men of American Society of Plant Biologists (ASPB) did a bold step towards ensuring research integrity in plant sciences. Fraud is no more, honesty restored, science and not Photoshop shall serve the greater good of agriculture and feed the humanity from now on. All this will become reality because Leonid Schneider has been blocked on Twitter by ASPB and their community site Plantae. Well, transiently, but still. It might have been a warning to me!
It’s not like I didn’t deserve getting blocked. Disinformation flow had to be stopped, about some shenanigans going on at ASPB’s flagship journal The Plant Cell, responsible for which were their own editors, including the newly appointed Editor-in-Chief Sally Assmann. These are the members and top-authors of ASPB whom I must thank for my tough but fair Twitter lesson:
One of these brave ASPB heroes whose honour needed defending is the gentleman I marked with a frame: Gary Stacey , Endowed Professor of Soybean Biotechnology at University of Missouri in USA. He is also ASPB member and one of recognised “ASPB top authors”. You can imagine the industrial scale cash which the soy and legume expert Stacey receives from agricultural multinationals, in recognition for his outstanding research which is now being questioned. And I trolled him on Twitter.
The Stacey affair started rather small, with relatively minor issues in his papers. It soon escalated as readers began to send me evidence of data manipulations for which no innocent explanations were possible, some of it was posted also on PubPeer. Yet it seems, Stacey’s employer the University of Missouri closed the case already. My attempt to submit additional evidence was not answered. Stacey declared on Twitter that the university attested him a “clean bill of health”, the university did not deny that.
For example, look what Stacey lab in Missouri (which the boss runs together with his wife and former PhD student, Minviluz Stacey) published:
This Photoshop hack job saw the light of the day almost 20 years ago, as Figure 10 in this paper:
R. Bradley Day, Crystal B. McAlvin, John T. Loh, Roxanne L. Denny, Todd C. Wood, Nevin D. Young, Gary Stacey
Differential expression of two soybean apyrases, one of which is an early nodulin
Molecular plant-microbe interactions MPMI (2000) doi: 10.1094/mpmi.2000.13.10.1053
That Frankenstein Figure 10 apparently was already used as Figure 9D in same publication, and still it is not the only fabricated dataset there, look at Figure 7C:
Stacey’s reply to this evidence of blatant data fakery in his old paper was spot on. I need to get a life, apparently. Maybe. Or maybe Stacey’s own life ambition as plant researcher has always been to compete with Olivier Voinnet?
Technology has developed since, Adobe Photoshop kept issuing new editions and people became more skilled in the high art of illusions of scientific reality. The following gel collage looks more professional, it appeared in the ASPB journal Plant Physiology labelled in caps as “BREAKTHROUGH TECHNOLOGIES”, probably ironically.
Yaya Cui, Shyam Barampuram, Minviluz G. Stacey, C. Nathan Hancock, Seth Findley, Melanie Mathieu, Zhanyuan Zhang, Wayne A. Parrott, Gary Stacey
Tnt1 Retrotransposon Mutagenesis: A Tool for Soybean Functional Genomics
Plant Physiology (2013). DOI: 10.1104/pp.112.205369
In the next years, new Photoshop mastership levels were achieved in scientific publications. It seems that cheaters of today do not just manipulate figures for publication, they are even capable of fabricating pictures of original raw data in Photoshop, in case a journal editor or PubPeer commenter starts asking questions. Maybe this is why the journal eLife is still trying to figure out why the unrelated loading controls (shown below) look so strangely near-identical.
Yangrong Cao, Yan Liang, Kiwamu Tanak , Cuong T Nguyen, Robert P Jedrzejczak, Andrzej Joachimiak, Gary Stacey
The kinase LYK5 is a major chitin receptor in Arabidopsis and forms a chitin-induced complex with related kinase CERK1
eLife (2014) doi: 10.7554/elife.03766
That Cao et al eLife 2014 paper contains also a spliced-on gel lane in the Figure 5B (below, on the right), just as another paper from same time period (Garcia et al, Planta 2013) does, shown on the left. What was wrong with the original gel band so it needed replacing? Another stealthily spliced gel was spotted in Liao et al, New Phytologist 2017, but there Stacey explained on PubPeer that original data from less than 2 years ago proved unavailable, so he repeated the experiments to verify the result. Ho-hum.
The left-hand image from Garcia et al Planta 2013 is not on Pubpeer, only on Twitter, where Stacey stopped replying. But regarding the Cao et al eLife 2014 paper, he took a very bold approach of denial:
“We were made aware of the above comments on this paper via the eLife editors. In their email to us, +they already dismissed the comments on Fig. 6 & 7 as unfounded and, hence, we will not comment further. However, the eLife editors asked us to provide the original blot for Fig. 5.”
Stacey then provided some faint low-resolution “scan” purportedly showing the original gel for Figure 5. He then insinuated that the splicing he was accused of was actually an artefact introduced by the publisher when his original low-resolution figures were up-scaled to high-resolution when published online. Stacey’s conclusion:
“As one can see, there was no fudging of the figure as suggested.”
“The eLife editors have informed us that they are satisfied with our response so we now consider the issue closed.”
As I learned, eLife is still investigating, the issue is actually not closed. But the good news for Stacey might be that a potentially more problematic paper of his in Nature Communications is likely to be safe from even the most modest attempts at correction. That journal namely proved its new stance on research integrity by having refused doing anything at all about a manipulated paper by Karl Lenhard Rudolph, despite announcements from the German Leibniz Society and the DFG. Good for Rudolph, good for Stacey, heck, that sounds like the Nature Publishing Group is seriously competing with Cell Press.
This is Stacey’s paper, from 2017 and now it gets even weirder.
Dongqin Chen, Yangrong Cao, Hong Li, Daewon Kim, Nagib Ahsan, Jay Thelen, Gary Stacey
Extracellular ATP elicits DORN1-mediated RBOHD phosphorylation to regulate stomatal aperture
Nature Communications (2017) doi: 10.1038/s41467-017-02340-3
The PubPeer thread started with a luceferase luminescence figure which Stacey lab managed to detect with a fluorescence camera settings. It soon became much more exciting. Two issues were flagged: first, in Figure 2b a set of two HA immunoblot bands seems to rather neatly match its next neighbours on the same gel, while in the panel below, Myc-bands look suspiciously like two different exposures of same set of 3 bands, only horizontally flipped.
The other issue was that two different immunoblots, probed with same antibodies but with different samples loaded, not only suggested overlapping band shapes, even background splotches looked same. These splotches and dirt artefacts can happen during protein transfer from gel to western blot membrane, and if they happen, they are unique. All in all, it was rather obvious the images showed different exposures of the same blot membrane probed with GST antibody. A minor mistake of oversight?
Stacey’s replies on PubPeer however went into a different direction. Instead of taking an easy way out and admitting an error, he posted some raw data which he expected to prove all concerns wrong.
The raw data for Figure 2 reminds one of a different case, where another scientist, Timothy Chan of Memorial Sloan Kettering Cancer Center, uploaded “raw” data to PubPeer, yet the two bands still overlapped perfectly, something which cannot be explained by chance, only by application of Infinite Improbability Drive. It seems both Chan and Stacey labs are constantly struggling with western blots turning into sperm whales and bowls of petunias. There are of course alternative explanations, and it doesn’t let the principal investigator look good while waving such “raw data” pictures about.
But what Stacey uploaded for Figure 3a, still looks exactly like two exposures of the same western blot. Yet instead of finally admitting a small mistake, he even doubled down with same pictures in darker, still denying they show the same blot, and added as precaution:
“the net effect is that none of this has any real impact on the overall conclusions of the paper.”
Eventually, Stacey announced to be doing some “inquiries” into the mystery of how the gel bands happen to turn out looking identical in his lab. In this regard, there is also this other fresh paper in the ASPB journal Plant Physiology which Stacey coauthored with colleagues from Washington State University:
Diwaker Tripathi, Tong Zhang, Abraham J. Koo, Gary Stacey, Kiwamu Tanaka
Extracellular ATP Acts on Jasmonate Signaling to Reinforce Plant Defense
Plant Phys (2018) doi: 10.1104/pp.17.01477
How did the gel band similarity there ever happen? Another magic coincidence? Or is this maybe the reason why the western blots in that paper lack all noise or background, to hide traces of data manipulation? Why does a journal accept figures of western blots looking like this, in 2018?
It is probably best to conclude this Stacey saga with a paper in Frontiers. It is also the freshest Stacey paper, merely a month old. If only Frontiers were able to muster enough decency and competence to ask for original data from their paying customers, and then act appropriately. Because in that paper, the soybean lab published some actual artwork.
Md Shakhawat Hossain, Nhung T. Hoang, Zhe Yan, Katalin Tóth, Blake C. Meyers, Gary Stacey
Characterization of the Spatial and Temporal Expression of Two Soybean miRNAs Identifies SCL6 as a Novel Regulator of Soybean Nodulation
Frontiers in Plant Science (2019) doi: 10.3389/fpls.2019.00475
The Photoshop art was spotted by none other than the famous image duplication detective Elisabeth Bik. Look what happened to the soy bean background:
It is not clear what the background beautifications in Photoshop were needed for, maybe to help Figure 1E looks less identical to the Figure 5G? Stacey remains silent on that PubPeer thread.
If you are interested to support my work, you can leave here a small tip of $5. Or several of small tips, just increase the amount as you like (2x=€10; 5x=€25). Your generous patronage of my journalism will be most appreciated!