Cardiff University misconduct investigation leads to Journal of Immunology retraction

The senior author of a Journal of Immunology paper has retracted it after a university investigation found that he had inappropriately manipulated images, Retraction Watch has learned.

Here’s the notice:

We wish to retract the article titled “Targeting Neural-Restrictive Silencer Factor Sensitizes Tumor Cells to Antibody-Based Cancer Immunotherapy In Vitro via Multiple Mechanisms” by Martin V. Kolev, Marieta M. Ruseva, B. Paul Morgan, and Rossen M. Donev, The Journal of Immunology, 2010, 184: 6035–6042.

Bands in Fig. 1A and 1B were pasted from multiple gels without indicating that this had been done. The last author, Rossen M. Donev, takes full responsibility for this action; the other authors were unaware of and had no part in the manipulation of the images. The findings and conclusions of the above article have been independently verified. However, due to the inappropriate manipulation of the data, we wish to retract the article.

A spokesperson for Cardiff University, where Donev had a position until recently, tells Retraction Watch:

Cases of poor data-handling are extremely rare at Cardiff University but the University takes a highly serious approach when such allegations are made. In this case, there was a full internal inquiry which established that one individual had inappropriately manipulated primary data in two figures in the paper. The individual was solely responsible for preparing these figures, and none of the other authors on the paper were aware of the manipulation.

The inquiry found that there had been no intention to mislead and subsequent repeats of the original experiments have shown that the paper’s conclusions remain sound. Nevertheless, in accordance with best academic practice, the individual responsible, who is no longer at Cardiff University, has agreed to the paper’s retraction from the journal.

The University is reminding staff of the need to observe appropriate methods when handling data, and is confident that the highest standards of practice will continue to be met.

The Medical Research Council, which funded the work, tells us:

Cardiff University informed the Medical Research Council (MRC) in September of the outcome of an investigation into inappropriate manipulation of data in the paper Targeting Neural-Restrictive Silencer Factor Sensitizes Tumor Cells to Antibody-Based Cancer Immunotherapy In Vitro via Multiple Mechanisms, and that Dr Donev would be writing to the Journal of Immunology to retract the paper.

The University also reported that they had independently repeated the experiments and had shown that the results could be reproduced.

The research was funded by the MRC; the MRC grant had already ended by the time we were informed of the irregularity.

The paper has been cited once, according to Thomson Scientific’s Web of Knowledge.

We’ve tried to reach Donev and the University of Swansea, where he has an academic appointment, for comment, and will update with anything we hear back.

Update, 12/20/11, 8 a.m. Eastern: Donev tells us by email how the figures came to be manipulated and that there will be no other retractions:

After the paper was published, an editor from the journal emailed me that some bands in Figure 1A,B have been pasted. I carefully investigated the issue and realised that I have overlooked and submitted Fig. 1A which has been prepared for an internal group discussion a few years before submitting the paper based on very preliminary data. I replied to the journal and admitted that this figure should not have been sent to the journal for consideration and that I have alternative data obtained at a later stage of the work. Regarding Fig. 1B, I did pasted the bands taken from multiple experiments and felt this fact was obvious from their different backgrounds, assuming that this fact should be clear to the readers similarly to a number of papers in which single boxed bands are presented from multiple gels. Unfortunately, I must have misled some readers and realised that preparing this panel in such a way and not mentioning this fact was a great mistake.     

A formal investigation was carried out by Cardiff University where the work was done. The investigation came to a conclusion that there had been no intent to deceive or perpetrate fraud. This was confirmed by independent verification of the findings and conclusions of the paper. The final report from the investigation was sent to the MRC. There will be no other retractions.

Hat tip: Axel Heiser

26 thoughts on “Cardiff University misconduct investigation leads to Journal of Immunology retraction”

  1. The “…bands were pasted from multiple gels..” but there was “..no intention to mislead.” I am not sure, firstly, what bands were jumbled around and pasted according to the figure (was it just a loading control or the actual protein[s] of interest being studied?) because the article is no longer accessible (just the retraction notice); and, secondly, whether you can rule out intention to mislead solely based on the retraction (based on reasoning I just stated).

    In other words, it’s kind of difficult to determine what bands were actually tampered with. If it were beta actin or GAPDH or something, I could reasonably assume that there may not an intention to mislead, but if it were the protein they were studying, then, well, that changes it.

    Of course, basing scientific misconduct rationalization on whether one is mislead or not, is not really the point, in my opinion.

  2. If the conclusions stand, is it truly productive to science to retract the paper? This looks more like a retraction because of breached decorum if on such grounds.

    I am totally against hiding the paper from public view. I am free to peruse the paper, am I not? What if I disagree with the retraction and want to base my research on some published aspect of the hidden paper?

    I do not think readers should be forced to accept any editorial decision like that. Science is made of scientists and what they think, and not of editorial and copyright traditions.

    1. Mr. M, as some others have already noted, it really is not that difficult to write a paper with completely bogus data (as in “never measured”) where the conclusions stand. I myself once had a paper essentially ready, all I lacked was the actual numbers (which *were* measured, I would like to stress!). The actual numbers did not alter my conclusions at all…so following your logic one could ask “why measure it at all?”

  3. If the figure was so badly botched as to require a retraction, then why bother including the bit about the results being “independently verified.” If the results are sound, then the results are sound. A correction with apology would be more appropriate. The way the authors have left it, there’s no way to check the veracity of their “independent verification,” and their work can no longer be cited.

    1. I don’t think it’s about the verification that the data are correct, in my opinion. It’s the means by which the data were displayed that makes the ‘purported’ data of the retracted paper unethical.

      In other words, I don’t believe the ends justify the mean in this case.

      1. A retraction indicates that the paper should never have been, and should not be, part of the scientific record. That’s a pretty strong position to take if most of the data, and the conclusions, are sound.

        It can put others working on similar topics in a tough position. Can you publish results that largely overlap with a retracted paper? I don’t know. It’s pretty awkward to cite a retracted paper for other publications and/or grants, even if you have reproduced and verified the results of the paper. And wouldn’t it be dishonest to not cite the work, especially if you’re building on its ideas?

        You don’t feel that a correction, fully explaining why the figure has to be corrected, isn’t equal in terms of its moral hazard? It leaves the scientific record correct and largely intact, while fully disclosing the sins of the paper.

  4. Donev seems to be in the habit of pasting images and letting the reader assume they’re all from the same gel…

    http://cancerres.aacrjournals.org/content/66/4/2451/F1.large.jpg
    http://cancerres.aacrjournals.org/content/66/4/2451/F4.large.jpg
    (the latter one is quite bad)

    Fig. 1A of PMID 11195781 also has seams in the lower panel. Panel B looks dodgy too. Same for Fig. 3 and Fig. 5 (possible reuse of bands between figures).

    Fig. 1B of PMID 17597212 has seams. Fig 2A and 2B use the same ladder/MWmarker bands on different images.

    Fig. 2A, lower panel (western blot) of PMID 18632654 contains dodgy seams, as does panel C…
    http://cancerres.aacrjournals.org/content/68/14/5979/F2.large.jpg

  5. Here we go yet again. I think us regulars here know that there is often a rather large fire somewhere with this kind of smoke. I feel so sorry for the poor old Western blot. Its name is being dragged through the dirt over and over again. It is such a powerful technique in the right hands, but is so easily manipulated – both before and after the film is developed. Loading controls themselves are useless really, at least without the corresponding ponceau stain. Even then, there are no guarantees it is actually the same blot. Essentially, there is no solution to this kind of activity. Just plain old trust between humans and the painful process of cleansing science of the “bad apples”.

    In response to those who often say “is this really worth a retraction”, it is worth remembering again that these manipulations are hardly ever isolated incidents. You can see examples of this all over this site. Investigators who do this tend to do it routinely unfortunately and one retraction often leads to many.

    As for the “the conclusions are still valid” argument, it makes no sense at all. There are no conclusions, the paper no longer exists. The arguments are invalid. Journals should not be allowed to print these statements. They are misleading and confusing at best. Replication carries little weight in my opinion. I have said before that most capable scientists can write a paper with no data at all and with almost complete confidence that it will be replicated.

    This is becoming a very depressing saga. Is the future of digital imaging in publications in jeopardy? Will we eventually have to submit raw films etc for an editor to scan and manipulate?

    1. “Essentially, there is no solution to this kind of activity. Just plain old trust between humans and the painful process of cleansing science of the “bad apples”.”

      This statement is true of much data in science, especially when the results cannot be presented as an image. Trust is essential in this line of work, but you only really know the truth of something by reproducing it yourself.

      “it is worth remembering again that these manipulations are hardly ever isolated incidents.”

      This particular sentiment bothers me quite a bit. It has the tinge of a witch-mob mentality.

      “As for the “the conclusions are still valid” argument, it makes no sense at all. There are no conclusions, the paper no longer exists. The arguments are invalid. Journals should not be allowed to print these statements. They are misleading and confusing at best.”

      I basically agree. If the figure is correctable with the same underlying results, then why not publish a correction? If it’s not correctable, then don’t say that the results still hold.

      “Replication carries little weight in my opinion. I have said before that most capable scientists can write a paper with no data at all and with almost complete confidence that it will be replicated.”

      Replication and independent verification is a cornerstone of science. As I’m sure you know, it is how many cases of fraud or simple contamination are identified: separate laboratories cannot reproduce the experiments.

      While, yes, most capable scientists can write papers with no data and expect the results to be replicable, these papers fall into three categories: (a) theoretical, (b) already known, or (c) trivial.

      In the first case, the “experimental data” is largely replaced by derivation, and the results can be quite striking and significant. This, incidentally, accounts for a sizable chunk of Physical Review Letters, the premier research journal in physics.

      In the remaining cases, the results generally aren’t publishable as research.

      So, what’s your point? Yes, we can all write papers about something that’s trivial, already known, or 99.9999% known, but either we call them reviews, or we don’t get them published.

      I don’t see how this invalidates the idea that a non-trivial result can (and should) be verified by independent experiment, or establishes how a retraction is necessarily warranted for a single, mis-prepared figure that can be reproduced. Publish a corrected figure with a new, equivalent gel, and demonstrate how the results can be reproduced.

      1. I thought that Dave meant that replication from the same group carries little/no weight. I agree that replication is the cornerstone of science. But replication (or so-called replication) from a group that has already shown to act unethically? Nope, not helpful.

      2. “how a retraction is necessarily warranted for a single, mis-prepared figure that can be reproduced. Publish a corrected figure with a new, equivalent gel, and demonstrate how the results can be reproduced”

        Let’s assume you and I compete with each other working on the same scientific problem. I cut corners by forging data and publish my results ahead of you. You then will have no choice but to abandon your work since the problem has apparently been solved by me (think of all the resources you have wasted and the disastrous consequences to, say, the PhD student involved). If a few years later my misconduct is discovered, my paper will still be OK as long as I can prove that I guessed its conclusions correctly. It makes no sense.

        1. @chirality, @Brad Casali:

          Look, I’m in complete agreement with you that fraud require retraction. The statement from Cardiff University had a rather important sentence in it: “The inquiry found that there had been no intention to mislead and subsequent repeats of the original experiments have shown that the paper’s conclusions remain sound.”

          Fraud requires an intent to deceive. If the institutional review is correct, then there was no intent to deceive, and thus no fraud. I have no reason to doubt Cardiff University’s institutional review process, although I am surprised that the corresponding author would, or even could, make this kind of mistake.

          So, taking findings of the institutional review at face value, I’m still left wondering: why not issue a correction? What it really comes down to is that I want to see the new data so I can draw my own conclusions, instead of relying on retraction statements like “The findings and conclusions of the above article have been independently verified.”

      3. I agree with many of your points, even regarding the witch mob mentality. There is a risk that this blog is heading in that direction, but perhaps that is what was needed?

        I just wanted to clarify my points on replication. What I meant was that I do not think that it is any kind of defense for a retraction, regardless of the reason for it. Of course, it is essential for science in general but just confuses the whole retraction message in my opinion.

    2. Dave,

      Never a truer word “Investigators who do this tend to do it routinely unfortunately and one retraction often leads to many”.

      It was only after looking through 3 or 4 Paus/Bulfone-Paus papers that the penny began to drop. One of those odd feelings. Either one is mad, or there had been systematic fakery.

      The Leibniz Society did alert me inadvertently that there were several papers in trouble.

      As you can see the commission contacted me and asked me to write to it. I did not interfere with any investigstion.

      By way of explanation, Nancy Rothwell is the president of Manchester Universtiy, where Paus/Bulfone-Paus are faculty:

      — On Mon, 2/8/10, Marco Berns wrote:

      From: Marco Berns
      Subject: science scandal Manchester university
      To: newsdesk@men-news.co.uk
      Date: Monday, 2 August, 2010, 7:17

      Subject: mail to Prof. Ernst Rietschel head of the Leibniz Gesellschaft

      —– Forwarded Message —-
      From: Marco Berns
      To: nancy.rothwell@manchester.ac.uk
      Cc: president@manchester.ac.uk; Paul.A.O’Neill@manchester.ac.uk; alan.north@manchester.ac.uk; colin.bailey@manchester.ac.uk; alistair.ulph@manchester.ac.uk; martin.humphries@manchester.ac.uk; rod.coombs@manchester.ac.uk; colin.sterling@manchester.ac.uk; albert.mcmenemy@manchester.ac.uk
      Sent: Sun, 1 August, 2010 13:52:44
      Subject: mail to Prof. Ernst Rietschel head of the Leibniz Gesellschaft

      Dear Nancy Rothwell,

      I think it will be your duty to inform the grant awarding authorities of the outcome of the investigation into publications of two of your faculty. Please see the message from Professor Werner Seeger at the bottom.

      The issue about the specific publication are about reuse of data and statistical errors which are too consistent.

      I do not think it is just two publications as professor Seeger would have written “another”, so at least three, maybe four or five, or more. The two members of your faculty have 34 publications together. It is not obvious what one of the members of your faculty has done to be on these publications (perhaps another issue you should look into as there are some relatively simple rules). It may be that the University of Manchester does not abide by these rules. I leave it to you.

      Perhaps you might ask R. Paus what he has done to be on these publications?

      Here are the overriding rules for authorship:
      http://www.icmje.org/ethical_1author.html

      I draw your attention to the positive things necessary for authorship:
      “Authorship credit should be based on 1) substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; 2) drafting the article or revising it critically for important intellectual content; and 3) final approval of the version to be published. Authors should meet conditions 1, 2, and 3.”

      and also what does not qualify as authorship:

      “Acquisition of funding, collection of data, or general supervision of the research group alone does not constitute authorship.”
      Do his contributions fulfill the requirements for publication in the field of biomedicine in all of his publications?

      Your two members of faculty may very well “blame downwards”, but they do have a history of exploiting people from Eastern Europe, e.g. paying them the least possible and then whipping them to get “results”. What did they expect from such behaviour?

      Yours sincerely,

      Marco Berns

      —– Forwarded Message —-
      From: Werner Seeger
      e-mail: werner.seeger@uglc.de
      To: marcoberns@yahoo.co.uk
      Sent: Mon, 26 July, 2010 14:49:06
      Subject: Your mail to Prof. Ernst Rietschel

      Dear Mr Marco Berns,

      Professor Ernst Rietschel has forwarded your e-mail message
      dated July 9th, 2010, to our commission which has been
      established to investigate allegations against some
      publications including the paper by Budagian et al. EMBO J
      24, 4260-4270, 2005. As chair of the commission I will
      greatly appreciate if you were willing to make contact with
      me so that we may clarify the specific points you have
      raised in your e-mail mentioned above.

      Sincerely yours,

      Prof. Dr. Werner Seeger

  6. Also take a look at this article in Molecular Immunology, particularly figure 4:

    The mouse complement regulator CD59b is significantly expressed only in testis and plays roles in sperm acrosome activation and motility
    Rossen M. Donev, Baalasubramanian Sivasankar, Masashi Mizuno, B. Paul Morgan

    http://www.sciencedirect.com/science/article/pii/S0161589007002313

    BTW, Fig 1B (pointed out by Virgilstar*) has duplications. Lane “negative+p53” is the same as CD59.2, complete with crack

    * http://cancerres.aacrjournals.org/content/66/4/2451/F1.large.jpg

  7. Yeah, Marco you have a point, I know the issue there lies on the author. However I think all truth must remain divulged. If the authors have a solid conclusion on ugly results they can always publish it in legitimate manner in lower-standards vehicles. Or prove their point or request others to do it.

    Still I stand by the assumption that a retracted paper ought not to be hidden from readers. The red RETRACTED alert seems to me a much more positive approach. And, of course, lengthy, clarifying explanations of why the paper was retracted. Let the readers decide.

  8. @sfs: In the case where the data weren’t manipulated, I agree with corrections. But once someone starts pasting, photoshopping, or otherwise doing unethical things, I think retracting the work is the best. For instance, that retraction notice for the paper at hand noted that the overall results were corroborated.

    So what? People can still make things up and guess in the right direction. How do we know that other figures (in this paper) weren’t also manipulated, but weren’t part of the investigation, or otherwise went undetected because they didn’t paste things that were noticeable (e.g.: making up numbers, percentages, or screwing with error bars)?

    Once one figure of a paper is shown to be falsified, then the rest, in my opinion, of the paper is questionable, and a retraction should occur.

    1. In reply to scifraudster July 23, 2012 at 11:24 am

      When the manipulation is in more than one or two papers who knows where it ends?
      Before RW it was almost imposible to shift any of it.

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.