Authors retract PNAS Epstein-Barr virus paper for “anomalous and duplicated” figures

pnas 2515PNAS has retracted a paper on the cancer-causing Epstein-Barr virus just two months after publication, in a notice that fingers a now-former graduate student for manipulating figures.

The paper tries to explain how Epstein-Barr virus blocks the immune system’s attempts to destroy it. According to the notice, the three “nonexperimentalist authors” – identified in the paper as two P.I.’s from University of Texas at Austin and one from the University of California, San Francisco – didn’t know the figures “were not reflective of original Northern blot and immunoblot data.”

That leaves UT Austin PhD student Jennifer Cox under the bus. Her LinkedIn says she pursued a PhD from 2010-2015, though it’s unclear if she’s received a degree. Cox’s name is at the top of P.I. Christopher Sullivan’s list of past lab members, and she’s the only one on the page whose name doesn’t hyperlink to additional information, such as a contact.

The school issued a press release about the study that quoted Cox, which has been removed from the UT site but is still available on Science Daily:

“While this work does not immediately identify new drugs, the fact that such different tumor viruses have converged on the same strategy makes this an exciting pursuit for future therapies against viral cancers.”

Here’s most of the notice for “Pan-viral-microRNA screening identifies interferon inhibition as a common function of diverse viruses”:

The authors wish to note the following: “After publication, we were alerted to anomalous and duplicated immunoblot bands in Figs. 3D and 4C, for which we immediately contacted PNAS. Further investigation revealed additional band irregularities in both the immunoblots and Northern blots presented, including those in Figs. 3E, 4D, 5 A, B, and D and Figs. S1 and S4B. When asked, the experimentalist could not provide key original non-computer-manipulated immunoblot and Northern blot data that were consistent with the experimental results being obtained as claimed in the paper. We note that the three nonexperimentalist authors were unaware that the data presented were not reflective of original Northern blot and immunoblot data. We continue to investigate how this occurred and whether any of the remaining data are credible. For these reasons, all conclusions in the work must be viewed with skepticism unless independently validated.

Accordingly, we are retracting this paper. We deeply regret publishing with these deficiencies and apologize to the readers for not catching these problems sooner, and for all the wasted effort this has caused.”

We contacted Cox via LinkedIn and emailed Sullivan and the journal. We’ll update if we hear back.

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.

8 thoughts on “Authors retract PNAS Epstein-Barr virus paper for “anomalous and duplicated” figures”

  1. If all three co-authors were nonexperimentalists, how could they take responsibility for the data as authors at the time of submission and who on earth supervised the student?

  2. I feel that senior authors based on my personal experience have the complete knowledge of experiments, data and interpretations. However, when caught like this blame the junior person and if awarded a Nobel Prize on research take the full credit. I think, a senior author owns the publication and should be considered responsible for inaccuracies, problems and fraud.

  3. The URL formed from Jennifer Cox’s name by analogy with the remaining links to lab members still leads to a page, without, however, any contact information—nothing but this picture of text and another photo of her and her dog.

  4. AshK made a strong point to state that the senior authors should take responsibility. The journal stated what the different authors have done. One of those helped in the ‘design of experiment’. The other provided ‘analytic tools’. (Whether these activities warrant authorship is a separate issue). The only other author who played a role is Christopher Sullivan, the PI, who was involved in the design and analysis of the experiment (!), and was involved in writing the paper. So he should take full responsibility, explain what happened and apologize. He does not only have the responsibility as senior author, but also as major contributor and – definitely – as principal investigator.

  5. To be honest I have never worked at any lab where the PI did not see any original data generated (that’s what weekly meetings are for). If I was the professor of this group I would be looking closely at any papers that my PI generated (with or without J.Cox on) as only one person being responsible for so many dubious figures is stretching credibility.

  6. In all of my relationships with colleagues, I have often disagreed with the methods they have used or the interpretation of their data, but I have never questioned the authenticity of their experiments. It is true that Chris Sullivan did not catch this before it was too late. It is also true that the editor and the reviewers also did not catch this, as I am sure that many readers did not catch it. We are all susceptible to being deceived because we generally trust that the data presented is authentic, if we didn’t, what would be the point of reading other peoples work in journals in the first place?
    I do not know Chris well (I met him in passing once), but many of my colleagues and mentors know him well and all of them hold him in high regard. I am positive that this retraction has been incredibly difficult for him.
    At the end of the day the responsibility rests, of course, on the shoulders of the PI, but is it that difficult to believe that he missed the issues that so many others missed as well?
    I hope that this experience in no way jades Chris Sullivan’s relationships with his current students and post-docs. I went to his talk at ASV 2014, and the enthusiasm that he exhibited when talking about the people in his lab was tangible.

  7. If you are quick to blame the PI (Chris Sullivan) for not spotting this sooner, consider the alternative. That would be a lab where the PI is standing behind everybody’s back all the time and does not trust you or any data you produce. I know I wouldn’t want to work in such an environment. A lot of life is based on trust, and unfortunately, every once in a while, a few individuals break it. This is not a reason to presume everyone guilty until proven innocent.
    By the way, it is certainly possible to circumvent even the most draconian PI’s, given ill will.

    Full disclosure: I know Chris, and he is undoubtedly one of the best scientists I have met.

Leave a 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.