Immunology paper retracted because “documents were not archived with due diligence”

A group of researchers from Austria, Canada, Germany, and the U.S. have retracted a 2008 paper in the Journal of Immunology after being unable to verify the contents of some key figures.

Here’s the notice:

We wish to retract the article titled “Transcription Factor Gfi1 Restricts B Cell-Mediated Autoimmunity” by Chozhavendan Rathinam, Hans Lassmann, Michael Mengel, and Christoph Klein, The Journal of Immunology, 2008, 181: 6222–6229.

Examination of original Western blot data revealed that primary Western blot documents were not archived with due diligence. As a consequence, doubts on proper representation and the identity of some control Western blot bands shown in Figs. 4E and 6C could not be unambiguously resolved.

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

Corresponding author Christoph Klein — who won a major prize from the American Society of Hematology last year — hasn’t yet responded to our request for comment. We’ll update with anything we learn.

Update, 11 a.m. Eastern, 9/27/12: Commenter vhedwig points out that this is the second retraction this year for two of these authors. The first was in PLoS ONE:

The authors wish to retract the article “Transcriptional repressor Gfi1 integrates cytokine-receptor signals controlling B-cell differentiation” by Chozhavendan Rathinam and Christoph Klein. PLoS ONE 2007, 2(3):e306.

After publication of the article, concerns were raised regarding the control bands in Figure 5A, and in particular, whether the bands represent total STAT5. Examination of the original Western blot data revealed that primary Western blot documents were not archived with due diligence. As a consequence, the doubts about the proper representation of the control bands in Figure 5A could not be unambiguously resolved. This course of action is in line with the recommendation issued by Hannover Medical School.


14 thoughts on “Immunology paper retracted because “documents were not archived with due diligence””

  1. A cynic would say that they Photoshopped their Western blots and then hid the source data to prevent others from detecting the fraud.

  2. The failure of proper archiving is a smoke screen – you can’t fail to archive something if it never existed in the first place! About 5 minutes of searching brings up the fact that the “Lyn” blot (lowest panel) in Figure 4 of the J Immunol paper is actually a re-use of an actin blot from Figure 6 of this 2005 paper in the journal Immunity…

    Ergo, the Lyn blot never existed. It wasn’t “mistakenly archived”, it was just another blot, deliberately re-labeled and used for a completely different experiment in the 2007 JI paper. This sort of s*** doesn’t happen by accident.

    Interestingly, the same authors have another retraction in PLoS One earlier this year, for what appear to be similar problems:

    The lead author (Rathinam) is VERY well published, almost suspiciously so… Out of 26 papers, 5 in PNAS, 5 in Nature family journals, 3 in Blood, 4 in Cell family journals. I wonder how many of those papers have problems too?

    1. “As a consequence, doubts on proper representation and the identity of some control Western blot bands shown in Figs. 4E and 6C could not be unambiguously resolved.”
      “Figure 4 of the J Immunol paper is actually a re-use of an actin blot from Figure 6 of this 2005 paper in the journal Immunity”
      The mystery solved! The authors probably forgot that they had re-used the blot otherwise they would have informed Hannover Medical School about it.

  3. I agree. The wording looks like an attempt to pretend that in their lab they once produced “original” and “primary” pieces of evidence which they somehow lost or mistakenly filed in a place where they can’t find them again. But it looks more like they didn’t produce originals and they resorted to re-labeling other people’s work.

    I wonder if researchers should start making their own “forgeries” in their own labs using some system that is guaranteed to work. They could re-label to their hearts’ content without getting caught. The danger seems to come when they re-publish a figure that has already been published. (poor attempt at emoticon for “tongue in cheek”) :-^)

    1. I am always totally amazed at labs that get caught like this.
      When I was doing westerns I was usually drowning in them, the idea I would need to reuse controls of all things from another paper is just unthinkable I mean do they only use their special good-luck break a leg 13 well comb for westerns intended for publications, so all their other trial blots are no good?

      Perhaps I could sell my blots on ebay – guaranteed unused.

  4. For anyone following paleo-climate science ‘auditor’ Steve McIntyre, this is a remarkable event. Many paleoclimatologists have refused to archive data and code, making a proper analysis of their work impossible. And unlike in this case, the editors of the relevant journals have, for the most part, not only refused to sanction such behavior, but they have positively supported it by refusing to take any action. Science and Nature, I’m looking at you.

    1. Can you clarify that point? There is an enormous archive of paleoclimatic data (here for example: ) that is freely accessible. I guess I’m a bit confused by the use of quotes around the word “auditor”, are you implying that Steve McIntyre is not an auditor and doesn’t share his data, or are you saying that his auditing has shown a lack of data sharing?

      1. I recommend McIntyre’s blog:, where you can peruse the details at your leisure.The link you share is irrelevant to specific published papers. McIntyre documents in bloody detail many cases in which he has requested data and code to replicate results and been told to eff himself. He has contacted editors in order to get their help and been told to ask the authors – who he has already said refuse to archive their data. That being said, some journals do now require that data be archived to back up papers. I’m quire sure that neither Nature nor Science do.

      2. markbul, you may be “sure” but you’re wrong. You can read Nature’s rules for data archiving and availability here, for example ( And since paleoreconstructions from a whole slew of studies broadly converge on a similar set of interpretations, and since paleodata is strongly archived as downwithtime points out, one wonders what the problem is.

        Most likely it’s with the “auditors”! Mr. McIntyre, for example, in one of his attempts to discredit the work of other scientists (McIntyre’s pretence that application of Group A’s methods produced strong spurious signals from red noise), used a “red noise” set that wasn’t red noise at all, but which contained some of the signal he was attempting to “magic”. McIntyre then discarded the 99% of the data that least supported his pretence before presenting his “analysis”.

        The particular problem with that approach (“auditing for reasons other than scientific”) is rather highlighted by the individuals whose careers have probably been most damaged by “following” McIntyre. Emboldened by McIntyre’s efforts, Drs. Wegman and Said set out to do their own trashing of a climate scientist but were undone by fraud resulting in retraction of one of their papers (

        I expect Wegman and Said will not be the last individuals to come unstuck by “following” Mr. McIntyre. Sadly unethical practices tend to beget further appalling behavior if left unchecked. Who audits the “auditors”!

    1. Why – did I set off your ‘ad hominem’ alarm? Or are the fundamental principles of openness and integrity in scientific publishing new to you?

      When I took Bio 101 lab, we were given a sheet listing the requirements for every experiment. Among them were that the methods section should give sufficient detail to replicate the work. In a paper consisting entirely of a statistical analysis of a dataset, the results cannot be verified without the original data and the code used.

      We know now that many paleo-climatatologists have refused to release either their data or their code. That means that there work is literally non-falsifiable. Thus, it is not scientific. And to be clear, if I collect my own similar data, and do the same analysis with my own software program, I am not replicating the work – so falsifiability remains impossible.

      Journals in many fields now require archiving of data and of code. Climate science is not one of them.

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.