Musical figures: PNAS paper corrected with version of “intentionally contrived and falsified” Nature figure

pnasOne of the two corrections recommended by a McGill committee for work by Maya Saleh and colleagues has appeared, in the Proceedings of the National Academy of Sciences (PNAS).

As we reported last month, the committee found that

two figures in [a] Nature paper had been “intentionally contrived and falsified.” One of those figures was duplicated in a PNAS paper, which also contained an image that  had incorrectly labeled some proteins.

The committee recommended that both of the papers be corrected, and the PNAS correction  for “Confinement of caspase-12 proteolytic activity to autoprocessing” reads as follows:

The authors note that Fig. 6 appeared incorrectly. There was an error in the alignment of the molecular mass markers, and minor adjustments have been made to the assignment of caspase-12 bands. The authors also note that the source of the anti-caspase-12 antibody for Fig. 6 was Sigma (clone 14F7). This error does not affect the conclusions of the article. The corrected figure and its corresponding legend appear below:

The figure — available here — appears to be a modified version of Figure 4C in a Nature paper by the same authors, one of the figures that McGill’s investigation committee said was “intentionally contrived and falsified.”

We’ll be keeping an eye on how the group corrects the Nature paper.

15 thoughts on “Musical figures: PNAS paper corrected with version of “intentionally contrived and falsified” Nature figure”

    1. Also the other papers:
      Caspase-12 modulates NOD signaling and regulates antimicrobial peptide production and mucosal immunity, published in Cell Host & Microbe in 2008
      Cellular inhibitors of apoptosis cIAP1 and cIAP2 are required for innate immunity signaling by the pattern recognition receptors NOD1 and NOD2, published in Immunity in 2009
      Control of intestinal homeostasis, colitis, and colitis-associated colorectal cancer by the inflammatory caspases. Immunity. 2010 Mar 26;32(3):367-78.

  1. I’m sorry, I call bullspit on the correction. The authors claim in the materials and methods to have generated their own caspase-12 antibody in rabbits, and the asterisk in the figure is an unknown band detected by pre-immune control serum. But now they say they used a commercial monoclonal antibody generated in rats. So what again is that asterisk’d band?

    1. I agree completely. It is also especially weird that the “band of unknown identity” that is detected in the total lysate lanes is also detected in the material immunoprecipitated with the FLAG antibody. Sure, in an unpurified lysate non-specific binding to detection antibodies is common, but it is quite a coincidence if this “unrelated” protein also binds anti-FLAG…

  2. Nature already have a track record in correcting papers that should be retracted (see other Retraction Watch megacorrections) but this one is significantly different because here there is a publicly released report describing an institutional investigation which found evidence of misconduct.

    Let’s keep this simple – this WAS clearly misconduct because according to McGill’s own report, two figures in the Nature paper were “intentionally contrived and falsified.” The fact that McGill were unable / unwilling to find an individual guilty (and therefore avoided the word ‘misconduct’) is simply irrelevant; the paper contains invalid results, the review process was subverted and the paper must be considered unreliable.

    As I pointed out in the previous article on this topic, Nature’s editorial (29th April 2010) on this leaves little room for doubt: ‘If an institution’s report concludes that misconduct occurred, we usually insist on a retraction — and will issue the retraction ourselves if the authors refuse to comply.’ (see

    Nevertheless I would bet a lot of money that Nature will go for the easy way out – accepting the correction rather than standing up to McGill and insisting on retraction. Surely no one will care about just another mishandled case of research fraud?

    Or will they? You only have one reputation and if you keep kicking it, it will eventually get damaged.

  3. To be clear, the corrected PNAS figure is not only a “modified version” of the Nature figure, but it appears that the same blot was used for completely different experimental conditions (this was discussed in detail after the first Retraction Watch posting on this topic).

    1. For the journal to accept this makes a farce out of the whole endeavour.

      In my limited experience the Editors of these journals, even the high-profile ones, have no ability or desire to understand what they are dealing with when encountering research fraud. They act as if all the issues were ‘accidental’ even in the face of incontrovertible evidence to the contrary.

      That’s why Nature are the test case here as they have ‘talked the talk’ in quite a big way, actually going to the trouble of laying out their policy for all to see. Their editorial says very much ‘We know best, here is what the rest of you should do.’ But I don’t think they are prepared to ‘walk the walk’.

  4. I think there’s something incongruous about the dates of the two problematic figures. The figure in the Nature paper, that was intentionally falsified “reported” on caspase,1,4&5 interaction with caspase-12 and was published in April 2006. This collage figure was clearly assembled from the PNAS figure that “reports” caspase1 interaction with wt and mutants of caspase-12. The PNAS figure was published in March 2008 and is according to the current correction a mislabelled original. I do not understand how people’s minds work in these situations, but I would argue that it is more in line with human nature that the original experiment was in fact as reported in the first paper (ie caspase 1,4,5 interaction with caspase-12) but manipulated to fit the authors “requirements” and was subsequently recycled in the later paper. Ie it is more dishonest to completely chop up and relabel a figure than it is to duplicate a couple of control lanes and the loading samples.

    I cannot ascertain from the previous report on Retraction Watch which original auto-rads could not be found and I guess we will never know. However it is clear that this correction is dishonest in failing to draw the readers attention to the fact that the PNAS figure was previously published, in a manipulated form, in Nature. I believe the editors at PNAS should be informed and asked to respond. Is there a possibility for a RW letter to ask for an official response?

    1. A good idea but I would be astonished if there was any meaningful response. I’ve been through this before with a Nature journal and the outcome is predictable.

      The first hurdle one would encounter would be intense apathy – your first letter will routinely be ignored. If you persist and make a second enquiry and start raising your ‘cybervoice’ you will discover that the person who should be dealing with it will have moved to another department and so it will be referred to a stand-in who is new to the position and unfamiliar with the details of the case. They will promise to discuss it in a meeting and another three months will pass.

      If you persist still further eventually you might get a response – somehow the journal can’t quite bring itself to ignore you completely when they have said they will deal with the enquiry – in their minds this could be worse than the scientific idiocy of their correction. If they eventually do provide a response, this is certain to support the original decision in a very passive and indirect way – a massive fudge basically. For example they might say that no one is really using this technique any more so the paper is of less relevance than it was when first published. And they will again refer to the committee’s recommendation of correction (despite the huge conflict of interest) and the lack of the word ‘misconduct’ – basically a whole series of little things that they can pretend justifies their decision. Their hope is that you will now just go away. And this works because in the end few people want to take it further and expose their own name in order to attack both a group of authors and the journal that has published their correction.

      Journals’ procedures for dealing with this in the end resemble cooked spaghetti. Occasionally you get a surprise – JBC has a good track record in retracting articles against the authors’ wishes (and they never get any comeback from the authors). But for some reason the general rule is that editors who review hundreds of articles per week and take tough decisions on a daily basis suddenly become shrinking violets when it comes to research misconduct.

      1. All valid points amw, and I have similar experiences to you in dealing with editors, but in this case the issue is not about the science etc but something more tangible.

        In my opinion, the correction commits the offence of “republication” because the figure in the Nature paper which is “identical” has not yet been retracted/corrected. Furthermore on this occasion they cannot claim that they had forgotten/overlooked the Nature paper because this PNAS correction is driven by the McGill report highlighting the problems in both papers.

        I am happy to write on my own behalf but I believe that if RW write (and this case as I have stated it appears to me to be within the remit of RW) then it is more public and official and much harder for the Editors to shuffle around until it dies a quiet death. I am happy to draft such a letter if there is any will.

        1. A good point – in summary you are saying the correction can be demonstrated to be clearly misleading because it avoids mention of the more fundamental fact that the image was duplicated from a different figure in a separate paper.

          The journal could claim, as things stand, that it did not know about the link to the Nature paper, but they should have made more enquiries before simply publishing the correction and trusting the McGill committee’s recommendation which is clearly flawed.

          It would certainly be worth trying but I still think the only way to make journal editors react is if they fear public exposure, and that is raising the stakes….

          There isn’t much middle ground.

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.