CNRS hits back at the stream of misconduct evidence

The French scientific society CNRS, a huge country-wide network of research institutes and one of the most influential science institutions in Europe, had enough of me and my reporting. They now blocked me on Twitter, so I don’t spoil their celebration of a book by their senior director and chief of research integrity, Catherine Jessus. The book is titled “Étonnant vivant”, which translated roughly means “Amazing ways to improve your Life science publications with just a little bit of photoshop”. I previously reported about some examples of this art in Jessus publications, which CNRS did not really appreciate.

And of course CNRS did not take it lightly when my past reporting forced them to deal with the Olivier Voinnet affair, whose investigator was Jessus. Neither did they like my writings about another misconduct scandal she had very successfully kept under wraps, that around the former Voinnet postdoc Charles-Henri Lecellier, now CNRS group leader in Montpellier. What finally took the biscuit was my constant leaking of misconduct evidence and of internal information from the former Voinnet institute in Strasbourg,  Institut de Biologie Moléculaire des Plantes (IBMP), or as I keep proposing to rebrand it, The Olivier Voinnet Institute for Research Integrity in Plant Sciences. CNRS initially tried it nicely, by sending the IBMP deputy director Jean-Luc Evrard to call me “baby”, then “idiot”, while instructing all other institute employees never to communicate with me, and eventually CNRS simply blocked me on Twitter.

On this occasion, I will now bring more evidence on data rigging inside IBMP. Before dumping new evidence about the IBMP director Laurence Maréchal-Drouard, let us start with something traditional and well known, namely Voinnet and his sidekick Patrice Dunoyer, whose IBMP lab was supposed to be dissolved, but maybe he managed to save himself with a recent publication in Nature Plants.

Voinnet’s Sidekick

Dunoyer was declared reformed, his papers are most welcomed by scientific journals now. In fact, even COPE, the organisation of scholarly publishers, for some reason revered as the highest authority on research integrity, indicated in an article in The Scientist that any blacklisting of cheating scientists like Dunoyer was unethical:

“Charon Pierson, the secretary and interim treasurer of the Committee on Publication Ethics (COPE), says that COPE does not agree with the practice of journals placing temporary bans on authors who had committed misconduct. “I would say that the only responsibility of the journal is to scrutinize manuscripts,” she adds. “To deal with the scientists themselves—that’s the realm of the institutions, the laboratories, the funding agencies, the governments, all of those pieces of the puzzle.”

CNRS and their investigator Jessus dealt with Dunoyer as harsh as his excessive data manipulations made necessary: he got suspended for one whole month. Harsh, I know, but it had to be done. Yet Dunoyer never even pretended to regret anything. He never admitted any wilful data manipulation, at least not in public, and he certainly never came clean about all the data manipulations in his papers. He admitted only those manipulations which were already evidenced on PubPeer, yet he was forgiven nevertheless. So let us help him by showing some of those bits he forgot. Following is the evidence a reader of my site forwarded to me. How much more are the tenured IBMP scientist Dunoyer and his ex-boss, the ETH professor Voinnet, hiding, while being backslapped by the scientific community for their newly found research integrity?

Here the paper Dunoyer et al, Nature Genetics 2007, never corrected despite 3 year-old PubPeer evidence. Now also this, an apparently digitally duplicated gel band:

Or how about this, in Dunoyer papers Navarro et al Science 2006 and Dunyoer et al Nature Genetics 2006, both corrected (here and here), but never flagged on PubPeer? Here Voinnet’s notorious multi-purpose loading control library makes its return:

But it is of course just those silly loading controls, which nobody needs anyway, right? Not really, loading controls are probably the most important part of a gel-based figure, I wrote about those here. So here is more of the same, never before noticed in Dunoyer et al, Nature Genetics 2006, on top of the ancient PubPeer evidence:

 

 

The director and her PhD students

If you believe CNRS will finally get rid of cheating Dunoyer after that, don’t hold your breath. His home institute IBMP has a certain tradition in this regard. For example, Dunoyer’s IBMP director Drouard has a PubPeer list longer than his. So now, here comes the evidence yet another reader of my site forwarded to me, it is mostly about one apparently heavily manipulated figure and its many incarnations. The primary focus is on this Droaurd paper I wrote previously about, and which was afterwards also flagged on Pubpeer:

Thalia Salinas, Samira El Farouk-Ameqrane, Elodie Ubrig, Claude Sauter, Anne-Marie Duchêne, Laurence Maréchal-Drouard

Molecular basis for the differential interaction of plant mitochondrial VDAC proteins with tRNAs
Nucleic Acids Research (NAR), 2014. DOI: 10.1093/nar/gku728

The paper contains this Figure 1A, where the lane “W” seems to contain a background patch which appears no less than four times:

A similar figure appeared already in a 2009 PhD thesis by the second author of the NAR 2014 paper,  Samira El Farouk Ameqrane, done under the supervision of Drouard as “Directeur de thèse”. Here it is as Figure 33A, but the Coomassie blue protein staining panel “St” is different. In the Western Blot lane “W” however, similar background patches are highlighted, yet the pattern looks very different from that in Salinas et al NAR 2014 paper. The upper part is namely the same, but the lower part is different:

Yet the original of this image is to be found in the 2006 PhD thesis of the first author of the NAR 2014 paper, Thalia Salinas, also done under supervision of Drouard. here it is as Figure 35, red and blue arrows indicate duplicated patches:

Part of this figure was then published in Salinas et al PNAS 2006, as Figure 1A, sans the first lane containing offending background duplications:

Not all Coomassie blue staining by Salinas worked perfectly for the protein VDAC. Sometimes they needed some digital fixing, like here in her thesis, Figure 36 and reused in Figure 46. The red arrow indicated contrast breaks evidence an inappropriate digital manipulation of these two images, while the green boxes show an apparently duplicated and mirrored section of a gel:

 

The Salinas et al NAR 2014 paper contains another problematic figure with a history. Some splicing was already flagged on PubPeer for its Supplementary Figure 4B, indicated here with red dashed lines:

This figure was stems from the 2006 thesis by Salinas, Figure 43. Here it is side-by-side with that from Salinas et al NAR 2014. Of course the reuse of PhD thesis figures in a paper is nothing to criticize as such, except of the hidden splicing Salinas introduced into her NAR 2014 paper:

If one looks closely at this same Figure 43 from Salinas 2006 PhD thesis, one detects splicing (indicated by green lines), while some background patches appear suspiciously similar, highlighted with red and blue boxes:

There is another version of this figure, in the 2009 thesis by El-Farouk Ameqrane, present as Figures 45 and 46. It is a beauty, what with its apparently duplicated lanes and bits of background.

The 2009 PhD thesis by El Farouk-Ameqrane contains another problematic figure, where a gel lane was digitally duplicated, as indicated by red arrows. This figure however never made it into Salinas et al NAR 2014, the experiment was repeated and a new gel for Figure 1C was produced. Did Drouard maybe notice that the thesis Figure 36 was perhaps too plumply manipulated to be published in 2014?

 

It is worth remembering that Drouard bears the responsibility for the content of both dissertations, that of El-Farouk Ameqrane and of Salinas. The IBMP director was however not interested to see any of the evidence of data manipulation in those theses, in fact she never replies to my emails on principle. Her past PhD student Salinas is presently a technician in her own lab at IBMP, tenured at the University of Strasbourg and married to the IBMP group leader Philippe Giegé. A paper from this lab turned out to contain duplicated gels (Gutmann et al 2012, PubPeer evidence and Giegé’s response here).

Will the journal NAR do something? Not likely. First of all, Drouard is member of NAR editorial board. Patrice Dunoyer was welcomed with his publications this year not only at Nature Plants, but also at NAR, as I reported here. The epigenetics researcher Irina Stancheva, who has an impressive PubPeer record and was recently sacked for misconduct from the University of Edinburgh (see my report here), is still listed as NAR Executive Editor (Update 21.10: NAR now quietly removed her, here cached website). Also:  another NAR Executive Editor is RNA-interference specialist Eric Westhof, from IBMP’s neighbour Institut de biologie moléculaire et cellulaire (IBMC) in Strasbourg. According to this document, Westhoff even collaborates with Drouard directly.

We can safely guess however what CNRS will do, about their Drouard, Dunoyer, Jessus or even Susana Rivas cases. Exactly nothing at all, except maybe of instructing everyone never to read my site. Or else.

---