High-profile plant biologist Olivier Voinnet has been suspended for two years from the Centre national de la recherche scientifique (CNRS) following the results of two investigations that revealed a number of issues in his publication record.
An investigation at ETH Zurich found that the scientist “breached his duty of care in the handling of figures as well as in his supervisory duties as a research director.”
Another investigation at CNRS, where Voinnet also works as a researcher, also found “the existence of deliberate chart/diagram manipulations” — while not “fabrication,” these breaches “amount to scientific misconduct,” according to a press release about the report.
The ETH committee uncovered multiple errors, from the “mere confusion of correct and incorrect figures” to “the ‘beautification’ of figures,” but said that it was not scientific misconduct, according to ETH guidelines:
- …in the publications in which incorrect figures were used, the scientific conclusions are fully substantiated by the available raw data.
- Based on the findings of the commission of inquiry, the Executive Board of ETH Zurich has concluded that this is not a case of scientific misconduct as defined in ETH Zurich’s Rules of Procedure.
The ETH report, which only covered work conducted at ETH, found “errors of varying severity” in 20 of the 32 studies they looked into.
The committee made an important distinction for Voinnet that the studies had been “conducted properly” but “published incorrectly.”
In the publications produced at ETH Zurich, the commission found errors in a total of five publications, ranging from the “beautification” of figures to the mere confusion of correct and incorrect figures. Serious errors were identified in two of these five publications. In some cases, the figures had been edited for internal use and then published in error. The commission examined raw data and documentation from the experiments that were carried out; these were complete and correct and substantiated the scientific conclusions made in the publications. Therefore, the processing of the figures led to no scientific advantage. Based on the findings of the commission of inquiry (PDF, 139 KB), the Executive Board of ETH Zurich has concluded that this is not a case of scientific misconduct as defined by the Rules of Procedure.
However, the researcher has breached ETH Zurich’s binding Guidelines for Research Integrity and Good Scientific Practice at the ETH Zurich. The guidelines state that project managers should take an active role in the management and supervision of junior scientists as part of their research. In this regard, and in the handling of figures, Voinnet clearly breached his duty of care. For this conduct, Voinnet will receive an admonition from the president of ETH Zurich.
In a separate inquiry, the CNRS committee found that Voinnet’s mistakes constitute scientific misconduct, but made a similar distinction between Voinnet’s actions and outright fraud.
The commission of inquiry has established the existence of deliberate chart/diagram manipulations, in breach of the ethical standards applicable to the presentation of scientific results. This involved modifications and duplications of diagrams/charts or erroneous captions. Such inappropriate presentation of experimental data, however, does not amount to fabrication.
Naturally, these mistakes will still have a serious impact.
These violations, which do however amount to scientific misconduct, have tarnished the reputation of the CNRS and of research at large. Consequently, the CNRS President has conferred with the disciplinary committee and taken appropriate measures, in accordance with procedures applicable to civil servants.
A spokesperson for CNRS said that the committee isn’t required to publish the proceedings of inquires, but decided to post their decision on Voinnet with full names withheld. The statement, which the spokesperson sent us directly, spells out the consequences for Voinnet.
Mr. V, a 1st class research director is suspended for two years from the decision putting an end to his temporary assignment.
ETH said that Voinnet’s group will now work with “an external advisor”, and has been encouraged to adopt “the introduction of electronic laboratory notebooks in the group’s laboratories” to help track their work. Correction notices and retractions will be published for every “erroneous” study.
Professor Detlef Günther, Vice President for Research and Corporate Relations at ETH Zurich, said that Voinnet’s cooperation had an impact on their decision.
“It must be clear to all researchers at ETH that we do not tolerate conduct of this kind. Voinnet failed to exercise the same care in publishing the data as he did when conducting and validating the experiments.” In Voinnet’s case, the Executive Board takes into account the fact that the experiments were performed correctly and, therefore, that the scientific conclusions in the publications are not being called into question. In addition, Voinnet has cooperated with the investigation from the start, has made his own proposals for correcting the problematic conduct, seriously regrets the errors he made and has contributed to a resolution of the errors by supplying the associated raw data. ETH Vice President Günther explains the Executive Board’s position as follows: “For these reasons, the measures focus on facilitating the group’s continued research activities, but also on rectifying the obvious shortcomings in the handling of figures.
In a response included in the report, Voinnet said that he regrets “not exercising the necessary care,” and offered an apology to his colleagues.
Olivier Voinnet takes complete responsibility for the errors made in the publications, saying: “I very much regret not exercising the necessary care during the publication process, and I take complete responsibility for all errors. I sincerely apologise for the uncertainty and difficulty this has caused members of my group and my colleagues. In future, I will work with the utmost care when publishing data. I would like to express my deep gratitude to ETH Zurich for its comprehensive investigation of the accusations against me, and for supporting me despite these errors. The university has placed its trust in me and I will not let them down.” Voinnet has begun to submit corrigenda relating to publications with erroneous figures, some of which have already been accepted. Two publications have been retracted in recent months based on the commission’s recommendation.
We have only tracked down one retraction so far, in The Plant Cell, along with a recent Expression of Concern. We’ll update the story once we’ve identified the second retraction.
Problems for Voinnet originally began when he received “an anonymous email” pointing out problems with his work. Afterwards, his research was heavily criticized on PubPeer, and more than half a dozen corrections have already been published for figures created in his lab.
Voinnet is a highly distinguished scientist who won the Rössler Prize in 2013 and was awarded EMBO Gold Medal in 2009.
We’ve contacted Voinnet and ETH for statements and we’ll update with any reply.
Like Retraction Watch? Consider supporting our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, and sign up on our homepage for an email every time there’s a new post. Click here to review our Comments Policy.
There’s something wrong with the timeline…
Investigations into Voinnet’s work started officially in APRIL (see : http://www2.cnrs.fr/en/2549.htm).
However in the last statement report from the CNRS (http://retractionwatch.com/wp-content/uploads/2015/07/Decision_V.pdf), they said “[…]investigation committee considered in a report dated 4 FEBRUARY 2015 that there was an established breach of the regulations and proper practices[…]”.
Could someone explain this gap ?
The press release was issued on 9 April, but it leaves unspecified precisely when, “Faced with these allegations, the CNRS set up a commission of inquiry, made up of leading experts.” I assume therefore that the commission was already well in place before 4 February, but was first publicized on 9 April.
From the report:
As OV has himself stated in one of the documents submitted to the
commission: “It is almost as if, having generated the data and convinced ourselves of
their robustness, the steps involving their representation under the form of figures (for
publication or presentation) became somehow less important than the fact of having
generated the data per se and answered a scientific question”.
Just amazing!
So he instinctively knew the data were “robust” and somehow that justifies misleading everyone else about their actual robustness as established by their reproducibility or by performing properly controlled experiments and presenting these honestly for publication.
“The ETH report, which only covered work conducted at ETH, found “errors of varying severity” in 20 of the 32 studies they looked into.”
This is not true. Voinnet came to ETH Zürich in 2010. The report, however, covered papers from 1998-2013. See page 2 of the report:
“Based on these accusations, a first ETH-internal commission viewed the criticized data, interviewed OV, and concluded that there was a need to establish an independent commission. Accordingly, this new commission was asked to further investigate allegations of research misconduct by OV concerning approximately 30 papers listed on PubPeer. These papers cover an approximately 15-year period spanning from OV’s PhD thesis work (e.g. Cell 1998 95, 177-187) to recent papers on which he had sole or shared corresponding authorship (e.g. PLoS Pathogens 2013 9:e1003435).”
In fact, the disciplinary actions taken by the ETH referred only to the papers that originated from the ETH.
Thanks! We’ve added a strikethrough to the relevant text.
Voinnet’s problems reminded me of Ernst Haeckel, the first European scientist to understand and defend Darwin’s theory and a superb biologist of aquatic organisms, but who diddled with the illustrations of his embryological research, to make them “more convincing”. Haeckel was roundly criticized, and lost respect, but as far as I recall, did not lose a position, although from what I can understand, his manipulations were more serious than Voinnet’s Perhaps we have better safeguards today?
Haeckel’s illustrations were presumably drawings (or etchings, etc.), not photographic (or similar) records, and as such would have been in any case
‘humanly interpreted’ in the act (whether his or an assistant’s, perhaps a medical illustrator’s) of producing them. Depending on how and to what extent he “diddled with” them (did he remove details he thought were extraneous, or combine details he had seen at different times? did he actively misrepresent quantitative features such as the number of organelles [say] of a given type, or merely modify [apparently] qualitative features? etc.), I might be more or less tolerant of his ‘diddling’. That he “was roundly criticized, and lost respect” of course suggests that whatever he did it was outside the accepted norms—but if his critics had other axes to grind (perhaps because they didn’t understand Darwin’s theory and wanted to attack it, not defend it), that suggestion would lose some of its force.
What exactly does it mean to be suspended, I wonder. Not allowed to publish? Not allowed to contribute to or benefit from ongoing work? Not allowed to visit the workplace or interact with staff? Step down from student supervision? Not allowed to authorize spending of grant money or submit new applications?
And apparently, at ETH, what you actually publish doesn’t matter as long as the results are correct. A slap on the wrist and everything as normal.
And what about this?
https://pubpeer.com/publications/F5BA8FE3EF8860205332839D02BBDD#fb27806
It is suspension without pay, so: all of what you said. Given that OV is not currently employed by CNRS, as French blogger Sylvestre Huet noted, it means in practice he cannot come back there. (The two-year suspension would take effect when he would come back.)
Suspended = no salary. So it’s unlikely he will ever come back working at the CNRS.
It means not having a place to work and not drawing a salary… If you can read French, see this informative article: http://sciences.blogs.liberation.fr/home/2015/07/le-cnrs-exclut-olivier-voinnet.html In brief, it basically means that he cannot return to the CNRS.
Back-to-back editorials about the Olivier Voinnet case:
http://www.labtimes.org/editorial/e_623.lasso
http://www.labtimes.org/editorial/e_624.lasso
I suppose the bright side of all of this is that the data wasn’t “faked” (as they still had the original data and the ETH could verify this) – although I find the phrase “beautification of figures” somewhat alarming (and un-enlightening).
For guidelines on scientific integrity and what is construed as misconduct the Swiss National Science Foundation links to a document by the Swiss Academies of Arts and Sciences.
http://www.snf.ch/en/theSNSF/research-policies/scientific-integrity/Pages/default.aspx
Swiss Academies of Arts and Sciences: Integrity in scientific research: principles and procedures (PDF, 1.1 MB)
Take a look at section 4.2
Specifically, the following clauses arouse interest:
“Scientific misconduct can occur in all fields of research. That is:
– in the expert appraisal of applications for grants and of the results of
research which are submitted for publication;
– through retaliatory measures, taken openly or covertly, against persons
who make allegations on the basis of inside knowledge obtained in the
research institute or through the checking of scientific data (so-called
whistle-blowers).
– deliberate falsification of data, false presentation and deliberately misleading
processing of research results, arbitrary weighting of data:
– exclusion of data and findings without declaration and without justification
(falsification, manipulation);
– concealment of the sources of data;
– elimination of data and materials before expiry of the statutory period
of retention of records*;
– refusal to guarantee authorised third parties’ access to the data.
– negligent or intentional wrong assessment of projects, programmes or
manuscripts;”
* In Switzerland, what is this statuary period?
However, this document was published by the Swiss Academy of Medical Sciences (SAMS). How applicable is this document to plant science?
If it is part of the SNF reference on Scientific Integrity, than I believe it would be broadly applicable to all SNF funded research. Furthermore, SNF guidelines are adopted by many if not most research institutions in Switzerland.
It is part of the ETH guidelines as well. This is an excerpt from the document
http://www.ee.ethz.ch/content/dam/ethz/special-interest/itet/department/Studies/ETH_Research_Integrity_2011.pdf)
Article 1 Purpose
The purpose of these guidelines* is to define guiding principles for the planning, execution, presentation and assessment of research work.
* These Guidelines have been elaborated particularly on the basis of the following documents:
1. a) Swiss Academy of Medical Sciences: “Wissenschaftliche Integritaet: Grundsaetze und
Verfahrensregeln, 2008,” (www.samw.ch)
2. b) Swiss Academy of Sciences: Draft rules «Grundsaetze zur wissenschaftlichen
Integritaet und zum Umgang mit wissenschaftlichem Fehlverhalten» submitted at
the experts’ meeting „Wissenschaftliche Integritaet“ on 20 March 2007 in Bern; and
3. c) Department of Health & Human Services, Office of Research Integrity (ORI): “Introduction to the responsible conduct of research”, by Nicholas H. Steneck (2004), (http://ori.dhhs.gov/documents/rcrintro.pdf).
New images of a 2008 EMBO Journal paper are being queried at PubPeer:
https://pubpeer.com/publications/61D31CDF08455DC10CF1F25D6E6A7F#fb33456
http://imgur.com/0m00A1z
http://imgur.com/9azzrHp
A news story at Nature:
http://www.nature.com/news/leading-plant-biologist-found-to-have-committed-misconduct-1.17958
Report by Ruiz, Voinnet, and Baulcombe (1998) corrected: http://www.plantcell.org/content/early/2015/07/17/tpc.15.00595.short
Retraction #2: 2006 PNAS:
http://www.pnas.org/content/early/2015/07/24/1513620112
http://www.pnas.org/content/early/2015/07/24/1513620112.full.pdf
“”PLANT BIOLOGY Retraction for “RNA silencing of host transcripts by cauliflower mosaic virus requires coordinated action of the four Arabidopsis Dicer-like proteins,” by Guillaume Moissiard and Olivier Voinnet, which appeared in issue 51, December 19, 2006, of Proc Natl Acad Sci USA (103:19593–19598; first published December 12, 2006; 10.1073/pnas.0604627103).
The authors wish to note, “Recently, potential errors affecting some of our figures were brought to our attention. Through the inspection of the corresponding original raw data, we have confirmed there were errors in Figs. 1C and 2 B, Left, E, Top Left and Bottom, and F, Right. While we have also confirmed that Fig. 2 E, Top Right, and F, Left, are not erroneous, our own inspection of all the remaining figures revealed additional mounting mistakes.
“Although these errors could have been corrected with the original raw data and were recognized as inadvertent we, the authors, consider them too numerous and have thus decided to retract the study on this basis. We will seek to submit the corrected version of this article for peer review elsewhere.
“Olivier Voinnet, the corresponding author, takes full responsibility for the errors contained in the original publication. We deeply apologize for the inconvenience caused.”
Another correction for PNAS 1999 paper:
http://www.pnas.org/content/early/2015/07/24/1513950112
Correction for “Suppression of gene silencing: A general strategy used by diverse DNA and RNA viruses of plants,” by Olivier Voinnet, Yvonne M. Pinto, and David C. Baulcombe, which appeared in issue 24, November 23, 1999, of Proc Natl Acad Sci USA (96:14147–14152; first published November 23, 1999; 10.1073/pnas.96.24.14147).
“The authors note, “The rRNA loading panels in Figs. 1D and 2D were inadvertently duplicated by Olivier Voinnet. Part of the image in Fig. 2D was also used in Figs. 3C and 5C, as explained in the respective legends. The experiments for Figs. 2, 3, and 5 were all run on the same agarose gel and blotted onto the same filter, to allow their direct comparison through the common reference dilution series seen in Figs. 2D, 3C, and 5C. The samples in Fig. 1D were run and blotted independently and it is therefore likely that the rRNA loading image in Fig. 1D, but not that in Fig. 2D, is erroneous. We no longer have the original files used for this paper and we are not able to provide the correct rRNA loading control to Fig. 1D. We recognize, therefore, that the equal loading of the samples in the figure is now not supported by the presented data. However, we are confident that the conclusions from the figure about the geminiviral suppressor AC2 are correct, as that interpretation was later confirmed in multiple independent publications.”