Paper linking fecal transplants to obesity in rats retracted for faked data

Diabetes Cover

A paper linking the fecal microbiome to obesity has been retracted after it became clear that one of the co-authors faked some of the data.

The 2014 paper in Diabetes — which found that rats given fecal transplants from obese mice were more likely to become obese themselves if given a particular diet — was pulled after after an institutional investigation found a co-author guilty of falsifying data underlying one figure and fabricating the data of two others.

Co-author Yassine Sakar — formerly based at the French National Institute of Agronomic Research (INRA) in Paris, France — was found responsible for the misconduct. But an official from the institution said that some responsibility must also be shared by the corresponding author Mihai Covosa, who has since resigned from the institution.

Here’s the retraction notice:

The lead and corresponding authors wish to retract the above-cited article as an institutional investigation has identified that coauthor Yassine Sakar falsified the data used to produce Fig. 4B, fabricated Fig. 5A, and fabricated the data used to produce Fig. S3(C). The authors regret these errors and apologize for any inconvenience to the readers of the journal.

The paper, “Replication of Obesity and Associated Signaling Pathways Through Transfer of Microbiota From Obese-Prone Rats,” has been cited 23 times, according to Thomson Reuters Web of Science.

Some have taken to social media to point out that one retraction from a field does not mean the whole discipline is false:

Sakar, who left INRA in February 2014, before the investigation began, has also notched a corrigendum for issues with figures of a 2014 paper, “Impact of high-fat feeding on basic helix–loop–helix transcription factors controlling enteroendocrine cell differentiation,” published in the International Journal of Obesity. The paper has been cited six times, and its correction has been cited once. It reads:

The original version of this article contained some errors in the figures. In Figure 1b, the x-axis labels ‘Visceral’ and ‘Epididymal’ were transposed. In Figure 2c, the second western blot image on the right was inverted and the images for Pax6, FOXA1 and CRIF1 were incorrectly represented. In Figure 2d, the blot image for FOXA2 was incorrectly represented. In Figure 3 a typographical error was introduced to the scale on the y-axis. In Figure 5b the blot images for PYY in the small intestine and colon and for secretin in the small intestine were incorrectly represented.

The corrected article appears in this issue and the html and online pdf versions have also been amended.

Olivier Le Gall, deputy director general for scientific affairs at INRA, told us the retraction notice does not address the fact the investigation at INRA concluded that some responsibility must also be taken on the part of the corresponding author. He noted:

Specifically, the investigation established the responsibilities of Dr. Sakar in compiling the data and preparing the figures, and of the senior author Dr. Covasa as corresponding author, and as “the guarantor of this work and, as such, [having had] full access to all the data in the study and tak[ing] responsibility for the integrity of the data and the accuracy of the data analysis” (quoting the “Authors contribution” section of the article).

Le Gall added:

When Drs Covasa and Sakar were given the opportunity to clarify the issues raised by the committee and that are eventually listed in the retraction notice, they failed to do so.

In Spring 2015, some of the co-authors of the article requested that the paper be retracted, but the last and corresponding author Mihai Covosa — who is now based at the Western University of Health Sciences in Pomona, California — did not agree to the retraction, according to Le Gall. The institution, therefore, set up a panel of both INRA and non-INRA experts in scientific integrity to examine the original data, and found “clear evidence of misconduct of several kinds,” explained Le Gall. He gave us an example:

…[E]veryone can easily notice that in Figure 5A, the top right corner of the “CVOP” panel overlaps slightly with the bottom left corner of the “Control” panel: as a matter of fact, a closer examination of the figure and the originals shows that all three panels of Fig. 5A were cropped from one same and single photograph field, while the legend states that they are supposed to belong to three different animals.

Next, said Le Gall, INRA asked Covasa to request the retraction of the article and ask the co-authors if they agreed. But, he noted:

Dr Covasa repeatedly refused to ask retraction in their names while the conclusions of the investigation committee were so clear, in September 2015 INRA as an institution asked Diabetes to retract the article.

Le Gall said the request for retraction from INRA copied in Covasa, and to their knowledge, he did not respond. The notice could have been written differently if Covasa had asked his co-authors to retract, Le Gall said, but instead it

does not fully render the corresponding author’s reiterated reluctance to request retraction of the paper.

In August 2015, Covasa voluntarily resigned from his faculty position as a senior scientist at INRA, Le Gall said:

That is, in a matter of days, around the time between the rendering of the committee’s conclusions to the head of INRA and INRA’s official command that he retracted the paper, but in his [resignation] letter Dr Covasa did not specifically mention the ongoing process.

Regarding the International Journal of Obesity corrigendum, Le Gall said:

…the current status of this article is that it was the subject of a corrigendum in November 2014 and, before this, an erratum in April 2014 (with the same DOI as the corrigendum) so the current online version is actually the 3rd one.

We’ve reached out the journal’s editor-in-chief, K. Sreekumaran Nair, a clinician at the Mayo Clinic in Rochester, Minnesota. Nair referred us to the American Diabetes Association (ADA) for a comment. Christian Kohler, staff liaison of the ADA’s panel on ethical scientific programs, said the panel reviewed and approved the retraction statement, but it was against the organization’s policy to comment publicly on details of such reviews.

We’ve also contacted Covasa for a comment, but could not find any contact details for Sakar. We’ll update the post with anything else we learn.

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

5 thoughts on “Paper linking fecal transplants to obesity in rats retracted for faked data”

  1. Questionable basis for the INRA institutional official’s statement about the retraction and investigation (which found the second author had falsified the data) that, “the retraction notice does not address the fact the investigation at INRA concluded that some responsibility must also be taken on the part of the corresponding author” [“as the guarantor of this work and, as such, [having had] full access to all the data in the study and tak[ing] responsibility for the integrity of the data and the accuracy of the data analysis” (quoting the “Authors contribution” section of the article)] . . . “given the opportunity to clarify the issues raised by the committee and that are eventually listed in the retraction notice, they failed to do so.”

    The paper has 7 authors (the one who allegedly falsified the data listed second), with the senior and corresponding author listed last, giving 4 different institutional affiliations: l’Institut National de la Recherche Agronomique in France; AgroParisTech in France; University of Suceava in Romania; and Western University of Health Sciences in the USA. The linked biosketch from WUHS https://www.westernu.edu/stp/bios.php?bio=mcovasa) indicates a WUHS faculty appointment since 2010 (and others in USA since 1998), but does not mention any INRA affiliation (although the biosketch does cite the questioned 2014 paper (fifth on the list of publications);

    Of course the senior author is “responsible” in general for the paper. But unless there is evidence the the senior author participated in the second author’s falsification, or knew that there was a likelihood that the second author was falsifying data and ignored it, it is inappropriate to imply that the senior author was responsible for that falsification (simply because the senior author reportedly did not take “the opportunity to clarify issues for the committee”). The retraction states, “The lead and corresponding authors wish to retract…”

    1. Alan, I do agree that it is probably possible to forge data in a way that makes it impossible even for your every day coworkers in the lab to detect it. But I very much doubt that this is usually the case. If a PI would really look at the raw data and lab notebooks every once in a while, fake data will become apparent most of the time. If you can’t asure that as a PI, you technically should not be senior author.
      Reality is completely different. In most cases, the lab boss puts pressure on everyone to get exciting data out and does not care much about how it came about. In fact, that is how it’s done even in other areas of life, not just science.
      If lab notes and raw data are faked so well that everything looks good and logical and chronological, an investigative committee will also have a hard time finding foul play.

      1. And to add to my point, as the senior author, I strongly believe that you need to have a look at the primary data before submitting the manuscript. So you ask your authors to supply you with the original blots and photographs etc. that were used in generating the figures in the paper.
        If someone used a blot or photograph more than once in the manuscript, like here in this case apparently in Fig5, there will be not enough original blots or fotos and the PI will find out right away. Reusing parts of blots or photographs and claiming different experimental conditions is the single most common offense and it is a lot more difficult to do if someone always takes a look at the full blot/photograph or whatever.

        1. Genetics — sure, but I have seen PIs who had looked at the raw blots immediately after they were produced each day, and then could not see changes made in the images by students or postdocs who submitted the final figures for their publication. But aggressive investigation committees have used sophisticated image analysis techniques on the PI’s papers and found those “hidden” falsifications or “beautifications” and then have “accused” the PI of being “responsible for the misconduct” — when the PI could not see these changes, nor understand why the student or postdoc would have made them, given the positive results on the original blots shown to the PI.

      2. Genetics – sure, but it happens to even the best labs with the most ethical and distinguished scientists in the country, who have not checked the raw data carefully, missed falsification of data in papers that they published, and then quickly retracted them. And they were praised for their openness, not tarnished nor held “responsible for misconduct” in their labs.

        Francis Collins (http://en.wikipedia.org/wiki/Francis_Collins) –University of Michigan Professor (who headed the NIH Human Genome Center, then the Institute, and since 2009 has been Director of the NIH — see the Hajra misconduct case at http://www.gpo.gov/fdsys/pkg/FR-1997-07-15/html/97-18453.htm
        (and Horace Judson’s 2004 book, “The Great Betrayal, Fraud in Science” on Pages 304-306)

        Leroy Hood (http://en.wikipedia.org/wiki/Leroy_Hood) — CalTech Professor, who won the Nobel Prize for his molecular biology work — see the Kumar and Urban misconduct cases at http://www.gpo.gov/fdsys/pkg/FR-1996-06-28/html/96-16561.htm and http://www.gpo.gov/fdsys/pkg/FR-1995-07-11/html/95-16961.htm
        (and CalTech RIO, David Goodstein’s 2010 book, On Fact and Fraud, on Pages 51-57)

Leave a Reply to Alan R. Price Cancel reply

Your email address will not be published. Required fields are marked *

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