A PNAS expression of concern appears — and so does its revealing backstory

pnas 1113When retraction notices and expressions of concern appear, particularly those that are opaque, we try our best to find out what’s behind them, whether it’s better explanations or the steps that led to moves. Today, we have one story in which we’ve been able to learn a lot more than usual.

In April, Bas van Steensel, Wendy Bickmore, Thomas Cremer, and Kerstin Bystricky sent a letter to about 80 leading labs in nuclear organization and steroid receptor biology. It began (we’ve added some relevant links):

On 3 September 2010, we sent a formal Letter-of-Concern to UCSD, HHMI, and the NIH Office of Research Integrity (ORI), with copies to the editors of Cell and Proc. Natl. Acad. Sci. USA. In that letter we expressed our concern that a team of UCSD/HHMI investigators, led by Dr. Xiang-Dong Fu and Dr. Michael Geoff Rosenfeld, may have presented inappropriately manipulated data in two scientific publications:

(1) Nuclear receptor-enhanced transcription requires motor- and LSD1-dependent gene networking in interchromatin granules. Nunez E, Kwon YS, Hutt KR, Hu Q, Cardamone MD, Ohgi KA, Garcia-Bassets I, Rose DW, Glass CK, Rosenfeld MG, Fu X D. Cell. 2008 Mar 21;132(6):996-1010.

(2) Enhancing nuclear receptor-induced transcription requires nuclear motor and LSD1-dependent gene networking in interchromatin granules. Hu Q, Kwon YS, Nunez E, Cardamone MD, Hutt KR, Ohgi KA, Garcia-Bassets I, Rose DW, Glass CK, Rosenfeld MG, Fu X D. Proc Natl Acad Sci USA. 2008 Dec 9;105(49):19199-19204.

The Cell paper was retracted in June 2008. The reason for the retraction, according to the authors, was a series of data duplications and transversions in the original dataset. The authors also stated in the Retraction that they continue to believe the central conclusions of the paper. One of us (BvS) expressed several serious concerns about massive data duplications and alterations to the authors around May 2008, which caused or contributed to the retraction. The Retraction text did not refer to
the mistakes as data fabrication.

Surprisingly, essentially the same story was published again in PNAS in December 2008 by the same 11 authors, under almost the same title. The paper was communicated to PNAS by Dr Rosenfeld – a National Academy member – within 10 days between submission and acceptance). In our Letter-of-Concern we provided multiple lines of evidence that suggest a substantial amount of data reported in these two publications may not be authentic. We requested that an independent committee should investigate this matter. Our extensively documented Letter-of-Concern (54 pages) can be downloaded here.

Officials at UCSD responded that they would investigate this matter. However, it took ~16 months, until 26th January 2012, before we were interviewed via telephone by an Investigation Committee convened by UCSD. On 29th November 2012 that Committee indicated to us that it had concluded its investigation. Their report was submitted to the UCSD Office of Research Affairs on 26th December 2012. The final report from UCSD was submitted to the NIH Office of Research Integrity (ORI) on 19th June 2013, i.e., with another ~6 months delay. Since then, UCSD and ORI have declined to comment on the case, despite requests from our end.

We therefore decided to re-engage with PNAS, the journal that had published the second paper and we sent them an expression of concern on 28th January 2014. To our surprise, PNAS informed us that they had recently printed two ‘corrections’ to the Hu et al. paper. These corrections included relabeling of figure legends to indicate that the data presented were from a completely different cell line than that indicated in the original manuscript – note our concerns about MCF7 cells below. The other correction involved the replacement of micrographs for which we had provided evidence in 2010 that the images could not be correct.

We are disturbed that PNAS should consider it acceptable to post such substantial data corrections to a manuscript that is under ORI investigation, and linked to a retracted Cell paper. Moreover, in 2010 the authors indicated to PNAS that the original microscopy images are no longer available. In our considered opinion, the numerous irregularities, duplications and mistakes associated with the publication either represent an unacceptable level of inadvertent errors, or suggest repeated data manipulation.

We are deeply concerned that – more than three years after the start of a formal investigation – the scientific community still has not been informed of the outcome of that investigation. We call upon UCSD and ORI to make the main findings of their investigation available without further delay, and we urge PNAS to publish an Expression of Concern, or retract the paper.

The letter — which we’ve made available here with the authors’ permission — goes into great detail about the problems they see in the paper. It also includes a timeline, which shows the stark difference between Cell‘s quite rapid response and PNAS’s not-so-rapid response.

Now, PNAS has issued an expression of concern about the paper:

The editors note that in January 2014, upon the authors’ request, PNAS published a correction to three figures in this article. Subsequent to the publication of the correction, the editors received concerns regarding images appearing in two of the corrected figures (Fig. 2B and Fig. S2) when compared with the data in the retracted Cell paper (1). In this regard, the authors note that the images in both the corrected Fig. 2B and Fig. S2 correspond to specific cells saved from large, multi-nuclei-containing photographic images. The authors note that this image-to-cell correspondence indicates that the incorrect cells shown in the retracted Cell paper resulted from errors in image processing. However, the original microscopy data is no longer available to further verify the accuracy of the images in question.

The paper has been cited 159 times, according to Thomson Scientific’s Web of Knowledge.

We’ve asked Rosenfeld if he has any response to the April letter, and will update with anything we hear.

19 thoughts on “A PNAS expression of concern appears — and so does its revealing backstory”

  1. I guess everybody should be concerned with the antiquated PNAS policy of letting members’ articles in so quickly and without any possibility of serious review.

    1. I know this is like comparing diamonds and moth-balls, but allow me to vent anyway. In this PNAS story, 4 individuals “sent a letter to about 80 leading labs in nuclear organization and steroid receptor biology”. They explained the problems and this resulted in an expression of concern. In parallel, I contacted 2000 out of 6000 attendees, including almost all key-note speakers, section editor board members and corresponding authors of posters and oral presentations at the 2014 International Horticultural Congress, to let them know that the “peer-reviewed” journal / proceedings book series that the International Society for Horticultural Science publishes, Acta Horticulturae, has potentially hundreds of cases of plagiarism, partial or full duplication, or just plain bad science. I then publically posted my e-mail here at RW [1]. I indicated some clear cases of two duplications, one which had been resolved with a retraction, and another that was due to the publisher’s error. I also indicated that at least serious case of apparent duplication (the entire paper was duplicated except for the figure; duplication = no cross-referencing by papers to indicate to readers that the other paper exists). These are all cases that involve the top echelon of the global horticultural community. One month later, and having received only one extremely defensive e-mail by a UC Davis professor, the horticultural community – most importantly – the management of the ISHS stay silent. In addition to this monumental effort which took weeks to achieve, having contacted all scientists individually about this problem, I have also, in recent months, been contacting over 3000 plant scientists about the problems in the wider plant science literature, particularly the ISHS Acta Horticulturae. I have indicated how essential post-publication peer review is, and have even published at least one paper (albeit in a Beall-classified “predatory” OA journal because at least 10 of the plant sciences journals rejected the paper based on the classical “out-of-scope” excuse) highlighting what I consider to be the 8 most concerning retractions in the plant sciences, excluding the Jorge Vivanco and Das scandals, which will be dealt with separately.

      So, from a grand total of about 5000 plant scientists, only 4 responses, and ZERO action. Either my approach is too intimidating, I was borne into the wrong sector of the scientific society that is more concerned with lavish symposia than with academic integrity, or – and this is my real fear – they simply just don’t care. As my former PI used to state when I used to talk to him about these serious errors in plant science academia and the literature: “I just want to retire in peace.” So, maybe that’s it. The vast majority don’t care; they just want to ride out their period on their earth only selfishly thinking that it revolves around their professional axis, and leave the errors for the next generation of plant scientists who will have a monumental task of cleaning up this mess the arrogant (or procrastinative, lazy, indifferent, etc.) generation of elites has left behind. I should note that the revolving door between many leading journals’ editor boards and these “grand” symposia, and the elite of the plant science community runs deep, probably because the global community is relatively small (even smaller when we consider horticulture, and larger when considering agriculture), but without a doubt incomparable to cancer, stem cell, cell (general), pharmacy, medical or so many other fields of study. So, to the critics of PNAS in this story and elsewhere, and to the critics of Nature and even some other top-level journals, actually, I am down-right jealous. Simply because at least these journals and scientists associated with them are responsive. I would die with pleasure if only 1% of plant scientists actually took note, and became responsive. I fight on, independent of the thumbs down and the criticisms, because we have this deep-rooted duty to science to call out what needs to be called out.

      [1] http://retractionwatch.com/2014/07/07/serbian-journal-lands-in-hot-water-after-challenge-on-24-hour-peer-review-that-cost-1785-euros/

  2. This is one of the examples that should be used to teach in ethics class – A senior author who is convinced of the conclusions (may be he knew right from the beginning of the project) despite evidence of data manipulation uses his clout as NAS member to publish though they had been forced to retract the same from Cell, a bunch of co-authors who are happy to bask in the glory of the publication with least concern about the authenticity of the data – first time can be excused since it is not possible for everyone to have seen the raw data but when a paper has been retracted for data issues still saying nothing or asking for verification(!!!!), random assignment of credit for the work – 1st author in Cell goes to 3rd in PNAS, its not the 2nd that becomes 1st but the 4th rises up when it is virtually the same story. Everything that is wrong about the scientific publishing today is present in this case

  3. This is not the first time this year PNAS has been criticised for its publishing methods regarding NAS members (see the “hobbit” story also covered by RW).
    Perhaps we should start thinking about re-classing this publication as a “vanity journal” ?

    1. So, PNAS is not all that it is cracked up to be. I wonder if less illustrious PIs befriend NAE/NAS members and get their papers published easily in PNAS by making them co-authors/communicators! A decision to publish in just 10 days for a paper of dubious value makes one wonder!

      1. Look carefully and you will find “less illustrious PIs” (friends/ former postdocs) using their relationship with NAS members to get the papers rejected from or unacceptable elsewhere being accepted at PNAS. Sometimes it is clear when the editor is ‘pre selected’ while other times you need to know the author and their association with the ‘editor’ to see through this blatant abuse of policies.

  4. It is apparent from reading the letters that Dr Fu starts to panic immediately after the first emails show up. He then (IMO) attempts to distance himself from his own data, claiming that they also noticed problems in an internal journal club a couple of weeks after publication. This seems a bit far fetched to me. First, rarely do departments review their own papers in journal clubs, at least in my experience. Second, if they were that noticeable at a journal club, how can one not detect these issues during preparation of the manuscript? Between the initial submission of a manuscript, the primary reviews, revisions and proof reading, the lead author should be so familiar with the material that he/she could probably recite the text verbatim like a script. It’s just not believable that these issues became apparent to the authors AFTER publication.

    What is absolutely remarkable to me is that the pair then appear to rely on Bas van Steensel to analyse their own raw data to detect where other problems are, and the authors seem to have no idea how to do this themselves. Quite shocking really, and there are only two explanations for this, none of which are great. They either are truly negligent in the prep of the paper and in the science, or there was more manipulation than they are letting on. One other point is that Dr Rosenfeld appears to be a peripheral figure in this from reading the correspondence, and as Al points out above, this is a lesson for PNAS.

    A disturbing case and I think there will be more to this story.

    1. Why do you say that Dr. Rosenfeld appears to be a peripheral figure? Both Nunez (1st author of the Cell paper) and Hu (1st author of the PNAS paper) are listed as affiliated to HHMI and therefore members of Dr. Rosenfeld’s lab.

    2. > One other point is that Dr Rosenfeld appears to be a peripheral figure in this

      “Peripheral figure” in what, exactly? In the email exchanges? Or in alleged data fabrication?

        1. Dr. Fu is indeed an independent PI and a co-corresponding author for these papers. But most of the authors seem to belong to Dr. Rosenfeld’s lab.

  5. The rule always is: publish paper, apply for grant. Thus I wonder, did Dr Fu get a nice fat grant off that Cell paper before it got retracted? And did he get another one from this zombie paper he was so far quite successful in keeping inside PNAS? Can someone check? Also, is Dr. Rosenfeld considering stepping down as NAS member, after this case of blatant membership abuse?

    1. oh dear, I just saw that Rosenfeld regularly publishes in Nature and Cell, often together with Fu. So, was this paper an unfortunate and unexplicable slip-up? Or did van Steensel et al uncover something bigger there?

    2. Thus I wonder, did Dr Fu get a nice fat grant off that Cell paper before it got retracted?

      Impossible to know. Dr Fu has 3 R01s. One he has had for 21 years, one for 15 years, and another for 7 years, He has been extensively funded for a long time.

        1. I think all one can say for certain is that his stellar publication record has certainly helped him get grants LOL! Whether one or two papers played a major role in him obtaining a grant is just difficult to pin down.

    3. As far as I can tell (by searching for “editorial expression of concern”) PNAS rarely publishes them, and then they are not followed up with a retraction. On several, it says they are awaiting the outcome of investigations,

      e.g. -from May 18, 2010, regarding PNAS 103:8354–8359, 2006
      “The matter has been forwarded to the Indian Institute of Chemical Biology (IICB) to investigate further. We are awaiting the findings of IICB to determine the appropriate next steps.”

      – and from March 19, 2010, regarding PNAS 101:8924-8929 2004 and PNAS 103: 2126-2131 2006
      “The University of Alabama at Birmingham committee has forwarded their findings to the US Office of Research Integrity (ORI). We are awaiting the findings of ORI to determine the appropriate next steps.”

      -and from June 3, 2008, regarding PNAS 98:8897–8902 2001, PNAS 99:15770–15775 2002, and PNAS 101:15736–15741 2004
      “We have been informed by the University of Perugia, Italy, of an ongoing review conducted by an inquiry committee at the university. We are awaiting the findings of the committee to determine the appropriate next steps.”

      Surely the investigations have been completed by now. PNAS should take the appropriate next steps.

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.