Research integrity

Remember Charles-Henri Lecellier?

Charles-Henri Lecellier is about to get promoted to CNRS research director 2nd class. Time to dig up old stories and let the ghosts rise to wash their dirty laundry.

The Raoultization of French science continues. The idea is apparently to establish a total kakistocracy of academic research, a glorious future where only fraudulent science, bunk and quackery is published, for the glory of France. This article is to celebrate another such promotion.

The greatest national achievements so far are:

  • IHU director Didier Raoult remains untouchable, in power, and in control of French COVID-19 narrative, his right-hand man Eric Chabriere is free to issue quack covidiocies, antisemitic insults and threats of violence to critics.
  • Also Guido Kroemer remains untouchable, maybe because disturbing him as the centre of data fudging universe would disturb the space-time continuum in France.
  • University of Nice rector Frederique Vidal remains France’s Minister for Research and Education, because she not only knows how to create imaginary research data, but also how to fight imaginary Islamo-Leftists.
  • Former CNRS president Anne Peyroche was absolved of all responsibility for proven research fraud and kept her job at CEA, French nuclear agency
  • Former CNRS chief biologist Catherine Jessus was whitewashed by CNRS and Academie des Sciences, then appointed as vice-president of Aviesan, a national pharma and biotech industry investment body.
  • et cetera

Jessus will play a role in the story I decided to dug up again. It is a story of the French geneticist Charles-Henri Lecellier, who is about to be promoted to the rank of CNRS research director 2nd class. Now, who is this gentleman?

Lecellier is the former postdoc of none other but the fallen star of plant sciences, Olivier Voinnet, the only high-ranking academic cheater whom French academia decided to find guilty of misconduct and to punish, and who I wrote quite a lot about (well, the French academia are back-pedalling on Voinnet’s guilt, too, now). Now, Lecellier was allegedly banned from Voinnet’s Strasbourg lab at the Institut de biologie moléculaire des plantes (IBMP, read here) because Voinnet did not trust Lecellier’s research results. Imagine this.

It seems Lecellier started to publish fudged data already during his PhD at CNRS Hôpital Saint-Louis in Paris, which, together with some falsified research in Voinnet lab, earned him tenure at CNRS. After the separation with Voinnet, Lecellier was appointed as group leader at the Institut de Génétique Moléculaire de Montpellier, where he continued his own line of data rigging, helped by his trusted friend from IBMP, Guillaume Vetter, now back in Strasbourg as the head of healthcare department at SEMIA, a regional startup investment non-profit agency. His achievements at SEMIA were celebrated in 2018 by Le Figaro.

It is this story of Lecellier and Vetter which I wish to call to memory, also because my original articles from 2015 for a German life science magazine were deleted together with their entire internet presence of their English-language magazine, Lab Times. There is only one short article rescued on internet archive, and below is the slightly shortened, edited and illustrated version of the main feature story I submitted to the now defunct Lab Times in 2015.

Vetter, Lecellier. Original photos: and

The Tale of Two Cheaters

This is a story of two friends with poor understanding of science integrity, and about the academic and industrial research networks which still protect them from facing the consequences of their research cheating. These two protagonists and long-term friends are the Frenchmen Charles-Henri Lecellier and Guillaume Vetter. Theirs is not a relationship of equals: though two years younger, Lecellier is clearly the dominant one and is greatly admired by his loyal follower Vetter, who also was said to have sought his mentor’s advice on the phone daily. But this is also a story featuring major institutional actors, namely the recently founded University of Luxembourg (Uni.Lu), the French Centre national de la recherche scientifique (CNRS), its subsidiary Institut de Génétique Moléculaire de Montpellier (IGMM), as well as the French biotech company Theradiag.

Lecellier used to be postdoc of the French plant scientist Olivier Voinnet, well-known for his research in plant immunity and the biology of small noncoding RNA. Voinnet was found to have committed misconduct and data manipulation throughout his entire academic career by the external investigation of his current and past employers, ETH Zurich and CNRS.

Lecellier, Dunoyer, … Saib, Voinnet, Science 2005. Never even corrected.

Since his time at Voinnet’s laboratory at the CNRS’ Institut de Biologie Moléculaire des Plantes (IBMP), Lecellier has focused his work on the studies of small noncoding RNA (precisely miRNA), but in mammalian cell models.  Several publications co-authored by Lecellier together with Voinnet are associated with the allegations of data manipulation; one was already retracted (Dunoyer et al., Plant Cell 16: 1235-1250). In his only email to me, dated April 7th, 2015, Lecellier wrote in this regard:

I am not aware neither implicated in the handling of that affair because I chose to get away from Olivier Voinnet 10 years ago. I would appreciate if you respect my past choices and avoid me today to pay the double penalty”.

Initially, this seemed to be a statement of an honest scientist who left Voinnet in protest against his data manipulation. In fact, the real reason for Lecellier’s break up with his mentor Voinnet was rather surprising. 

A former IBMP employee described Lecellier to me in this way:

Charles- Henri was extremely violent (verbally and physically) and arrogant. I never meet others as violent as him in a research laboratory. He was always with Olivier Voinnet”.

This witness never heard of anyone to have had left Voinnet’s team “on bad terms” and adds:  Lecellier was

one of the followers of Voinnet who had a totally inappropriate behaviour. It was tolerated by the hierarchy and even watched with amusement by other directors of research because Voinnet published in Science, Nature and Cell. But it was not fun for people of the institute who had to undergo all these inappropriate behaviour in everyday life at the lab”.

As I has learned from a first-hand source, Voinnet himself admitted that he did not trust his postdoc’s experimental data and removed Lecellier from his lab in 2006. As apparently unsackable French civil servant, Lecellier was transferred to the CNRS Institut de Génétique Moléculaire de Montpellier (IGMM). It seems therefore, Lecellier’s attitude to science integrity was even too much for Voinnet, as the proverb goes: “it takes one to know one”. Interestingly, the internet portal PubPeer easily provides evidence for Lecellier’s earlier forays into creative data presentation, as PhD student of the virologist Ali Saïb at the CNRS Hôpital Saint-Louis in Paris. The latter declared about his honour PhD student (Mention très honorable avec les félicitations du jury):

During his PhD, nothing could suggest any misconduct […] I’m not worried about the results since they have been repeated several times, including after the publication, and largelly discussed with all the members of the team. What remains totally unclear for me is why Lecellier used the same control blots for different figures for publication. I cannot know whether he did that intentionnaly or not.

PubPeer link

Yet Lecellier’s first-author publications on foamy viruses show strong evidence for manipulative re-use of blot images, in utterly different context and across separate publications (J Virol. 76(7):3388-94 and 76(14):7220-7). The journal’s Editor-in-Chief, Rozanne Sandri-Goldin, promised in an email to me:

given the public and sensitive nature of this case, we will follow our well-established process to investigate the allegations as they appear on PubPeer”.

The mistrust and conflict with Voinnet arose in 2006 at IBMP in Strasbourg over Lecellier’s first research project as independent scientist and his first collaboration with Vetter (who was previously employed at IBMP as well, in the laboratory of David Gilmer).* The data was produced by Vetter after he moved to Luxembourg, to work as postdoc at the CRP-Santé (now named Luxembourg Institute of Health), where he operated a DNA microarray platform managed by the Uni.Lu professor Evelyne Friederich. Despite the conflict with Voinnet and two journal rejections, the paper was eventually published in the journal Blood, with Lecellier’s wife Anna Saumet as first and Vetter as second author (Vol. 113(2):412-21). As a side note, Lecellier used to advertise his leadership skills by referring to Saumet as his “Supervisory Staff” in his CV.

The paper described the role of various miRNAs on the expression of cancer-relevant genes in leukemic cells carrying the common oncogenic mutation PML-RAR. Potentially, this study could be very important for clinical applications and patient therapies.  If it were reliable, that is. On PubPeer, once again a duplicated miRNA blot was flagged for that publication, while purportedly showing completely miRNAs under different experimental conditions.

Yet this is not all. I have obtained the manuscript version from March 2008 which was initially submitted by Lecellier to Blood. Here are the files for download:

I then compared it with the revised manuscript, which Lecellier resubmitted after peer review, and which was published by the journal.  

The exercise was similar to popular the children game of spotting the differences between two images and the reader is now invited to do the same. 

miRNAs repressed by PML-RAR control crucial cancer genes. Spot the differences!

There, it quickly became clear that certain values inside several gene expression analyses have been selectively altered, while other values of the same experiment as well as their error bars remained exactly the same. As aside, it is possible that also a certain “Q-method” was used, which Vetter allegedly introduced into Friederich lab and where the “worst” value out of a technical triplicate is always removed, while the mean value and standard deviation are calculated out of the two remaining numbers. Allegedly, an entire Master thesis got contaminated in this way and many more experiments in the Friederich lab. Nevertheless, neither editors nor the peer reviewers have noticed a thing.  Blood journal’s senior manager for peer review, Virginia Ramsey, has replied to my inquiries that these “have been forwarded to our Data Integrity Manager and are being processed accordingly”. 

miRNAs repressed by PML-RAR control crucial cancer genes. More differences.

The originally submitted Figure 3 contains real-time quantitative PCR (RT-qPCR) data for the miRNA expression in cells of 3 different leukemia patients. When compared to the Figure 4 in the published paper, one quickly notices that certain selected qRT-PCR values inside presented experiments have been changed. Why did the authors do this? Well, during the peer review, “Referee 2” has criticised the data from the patient samples as containing „significant variability”, mentioning that “primary APL cultures are unstable and often associated with significant number of dying cells“. To that, Lecellier has replied in the rebuttal letter:

The Figure 3 has been modified (new Figure 4). We hope that these modifications now improve the interpretation of the data. We acknowledge the existence of variability in our RT-qPCRs. As suggested by the referee, statistical analyses and additional RT-qPCRs have been performed“.

Yet only some of the values have been changed (often very dramatically), while others remained exactly the same. Precisely, the modifications were made to four out of five panels (Fig 4 C, E,  and G, but not D), and affected only two to four values out of the 15 presented in each panel. If Lecellier and Vetter really did perform “additional RT-qPCRs”, they did it for the selected values only, which is absolutely not acceptable in a properly controlled experiment. 

Similarly, a panel in the Figure 7 showing a luciferase assay for the miRNA effect on reporter gene expression was modified upon resubmission. It was extended by two new experimental conditions, with a key positive control significantly changed (the latter modification has made the impact of certain anti-miRNA treatments appear significantly greater). Once again, all other values (including their error bars) remained exactly as before. This practice by Lecellier and Vetter suggests to the least an inappropriate patchwork of intra-variable experiments, which prohibits any reliable comparison of values. Nevertheless, Lecellier insisted in his rebuttal letter that the added values can be directly compared to the previous ones, despite the profound intra-experimental variability. He only forgot to mention the introduced sudden dramatic change in the positive control. 

Having successfully published together, the Lecellier/Vetter duo moved on to a new project. It was largely performed in the Uni.Lu lab of the cell biologist and cytoskeleton specialist Friederich. I already reported online (June 10th, 2015) on the case of the resulting Oncogene publication (Vetter et al., Vol. 29, 4436-48) and the misconduct investigation by Uni.Lu. Vetter was found exclusively guilty by the external commission (as showed by the documents made publicly available on the website Figshare; also Retraction Watch has obtained a statement of Uni.Lu in this regard, posted on June 25th, 2015). 

The Oncogene paper postulated a role for a certain miRNA (miR-661) in the promotion of cancer cell invasiveness by inducing the epithelial to mesenchymal transition. The epithelial genes Nectin-1 and StarD1 were identified as key targets for miR-661, but the entire experimental data on Nectin-1 was proven by the investigative commission to have been manipulated by Vetter.

The paper was destined for retraction, which was strongly opposed by Catherine Jessus and CNRS, but also partly by Uni.Lu. The deal was therefore made to follow the commission’s alternative recommendation: to correct the paper despite its fraudulent data. Yet even there no agreement could be achieved, with Lecellier continuously arguing that there never has been any misconduct, but only “personal conflicts” directed at his innocent friend Vetter.

Letter by Catherine Jessus on behalf of CNRS (in French), stating a clear preference for a correction, but not a retraction, despite the results of the external investigation in Luxembourg

This entire sorry affair can summarized as such: hardly anyone wanted to accept Friederich’s and her research team’s accusations against Vetter, despite all presented evidence and institutional investigation. But why? 

The University of Luxembourg is very young (it was founded in 2003), but it already had a history of data manipulation. The first case was that of the German bioinformatician Carsten Carlberg. After an intervention from the Luxembourg funding agency, Fonds National de la Recherche (FNR), Carlberg was sacked from his professorship position at Uni.Lu for being involved in scientific misconduct, after two of his publications were retracted (Retraction Watch report from July 16th, 2011). That case resulted in Uni.Lu’s  first external misconduct investigation. Following this experience, an Ethics Review Panel was established at Uni.Lu.  Soon enough, it was needed: an assistant professor at the Life Science Research Unit (LSRU),  Eleonora Morga, was found to have manipulated data in a neurobiological study, which was later retracted by the journal Glia (Morga et al, Vol. 62(3):491; Retraction Watch report from January 1st, 2014). The paper’s last author was Morga’s former PhD advisor Paul Heuschling, Dean of the Science Faculty at Uni.Lu. Apparently, Heuschling, with the help of the then-Chancellor of Uni.Lu (now Vice-president for academic affairs) Eric Tschirhart, tried their best to prevent an external investigation and to avoid the retraction. This was despite the evidence presented by Morga’s own lab members that she manipulated data (in the Glia publication and elsewhere) and failed to provide her co-authors with access to primary data. Eventually, Morga was sacked. One of the reasons for the initial cover-up of her misconduct was money: FNR has a policy of demanding back their research funding connected to a publication retracted for misconduct. 

In fact, also in the case of the Vetter’s Oncogene paper, FNR was set to claim their money back, should it be retracted as well. Thus, not only CNRS was keen to avoid a retraction of Vetter’s heavily manipulated paper, but also the Luxembourg science directorate. 

Just when Vetter was supposed to be questioned by the investigators, he slipped away to his friend Lecellier in Montpellier. But their Oncogene paper was doomed. In October 2014, Uni.Lu leadership co-signed the official request to the journal to retract the paper.

Nothing happened, maybe also because CNRS as well as Tschirhart and even the Ethics Review Panel at Uni.Lu disagreed with a retraction. After my Lab Times report was online, the Nature Publishing Group has announced the retraction to appear for 29 June 2015. Instead, the back-and-forth negotiations about the wording of the retraction notice and the mention or non-mention of the external investigation and Vetter’s data manipulation continued, a retraction appeared only on 14 December 2015, its entire text was:

“At the request of the University of Luxembourg and following an external investigation, the Editor and Publisher have agreed to retract this paper owing to unreliable data.”

Unlike Uni.Lu, CNRS seemingly always stood united: on the side of Lecellier and Vetter. There are however some clues to certain financial interests. Four years ago, Lecellier founded together with IGMM the biotech company Prestizia, for the purpose of miRNA diagnostics and personalised medicine. It was later acquired by the diagnosis company Theradiag, which then went on to win the Worldwide Innovation Challenge from the President of France for its miRNA activities.

As IGMM proudly announces on its website, their staff scientist Lecellier supervises the collaboration with Theradiag on the HIV diagnostics research. Perhaps unsurprisingly, Vetter was given the position of the Head of miRNA Laboratory at Prestizia-Theradiag, as soon as he left Luxembourg and thus avoided facing the responsibility for his data manipulation. When contacted by email, Theradiag declined to comment on Vetter’s employment, but wished to know the names of those who provided me with this information. Thus, one could speculate that CNRS might have some high hopes for Lecellier and Vetter to procure some industrial cash through their involvement with biotech companies.

Vetter et al Oncogene 2010: Panel 4D, “-Tet” the dark image for nectin 1 shows a very different cell colony background as in SNAI1 and DAPI. Also, the SNAI1 image in “+Tet” lacks all background in Fig4D, unlike that in Fig4E

Of all these problematic Lecellier/Vetter papers, none has been corrected or retracted so far. CNRS remains silent. Uni.Lu now has another case of misconduct to deal with: the German neuroscientist Jens Schwamborn (Laborjournal online, 17.07.2015). Shortly after he was recruited as professor at Uni.Lu, Schwamborn was found to have committed misconduct and data manipulation in the two now retracted publications from his time as PhD student at the University of Münster.**

Except of the Oncogene retraction, and the retraction of his paper with Voinnet, Lecellier never had to correct or retract another paper. His partner Vetter is now in charge of healthcare startup investments in Alsace, maybe you need to supply your own misconduct findings if you wish to apply for funding. I was thinking of asking SEMIA if they know about Vetter’s past, but then… I learned that SEMIA president Lilla Merabet is accused of having embezzled €600k funding which she awarded to the startup of her partner. Vetter’s Luxembourg connection fits perfectly, SEMIA has partnered with that state investor to spend €350 million.

Meanwhile, Lecellier is on track to rise even higher in CNRS hierarchy, because his science is exactly of the quality France wants to have.

The article is an edited version of what was published in Lab Times in 2015.

* Right after the story was published in 2015, David Gilmer protested in an email to me: “I have to inform you that G. Vetter has never been my employee. Indeed, at this time (2003), I was assistant professor working in the same team as G. Vetter (PhD student). My responsibilities within the team did not consist of Vetter PhD supervision. However, G. Vetter worked with me to finalise a study. Therefore, I would appreciate my name removed from your article since at that time, this was not my laboratory nor my student.

** Read more about the case of Jens Schwamborn and the University of Münster here.


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!


0 comments on “Remember Charles-Henri Lecellier?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: