Nature pulls two papers on stem cells after investigation shows serious problems

Nature publishes retractions of two high-profile papers that claimed a major advance in the field of stem cells.

Science journal retracts paper on stem cell discovery.

The scientific journal Nature Wednesday retracted two stem cell papers that received national attention when they were published in January.

The paper by researchers from Harvard University and Japan’s RIKEN Institute described a new method of producing versatile stem cells without altering their DNA – a process that promised to make it easier to use stem cells in research and treatment.

Stem cell researchers immediately raised questions about these new cells, called STAP cells, and have tried unsuccessfully for months to reproduce the process of making the cells, as described by the papers.

For the inside scoop on these stories, I always head to Retraction Watch.

STAP stem cell papers officially retracted as Nature argues peer review couldn’t have detected fatal problems | Retraction Watch.

The editorial in Nature cites the problems with the paper that led to the retraction. But they say that there was no way that peer reviewers could have caught these error which only came out from an investigation into the study.

STAP retracted : Nature News & Comment.

We at Nature have examined the reports about the two papers from our referees and our own editorial records. Before publishing, we had checked that the results had been independently replicated in the laboratories of the co-authors, and we regret that we did not capture the authors’ assurances in the author-contributions statements.

We have concluded that we and the referees could not have detected the problems that fatally undermined the papers. The referees’ rigorous reports quite rightly took on trust what was presented in the papers.

The peer review process is broken and needs to be fixed. But for now, it’s still the best out there. Science is done by people. People make mistakes and screw up. Welcome to reality.

  4 comments for “Nature pulls two papers on stem cells after investigation shows serious problems

  1. Derek Freyberg
    July 2, 2014 at 11:57 PM

    The Fuji TV network’s “FCI Morning Eye” (a Japanese news broadcast rebroadcast in the US – here in the Bay Area by Tokyo TV) this morning showed a segment on the STAP cell controversy, including another press briefing by Dr. Obokata, the lead researcher on the retracted papers (and perhaps something else by RIKEN). Unfortunately, I couldn’t hear much of what was said, but it will be interesting to see if it is picked up by English-language broadcasters such as NHK World or the English-language press. The brief impression I got from the (Japanese-language) subtitles was that attempts to reproduce the experiments were ongoing and should be completed later this year, though I had the impression that there were some issues about who was doing the reproductions of the original experiments and whether they had the original protocol. Sorry not to be more complete, but I was in a restaurant, the volume on the TV was low, and my Japanese may not be up to the full details of this.

  2. Jon O
    July 3, 2014 at 9:00 AM

    Are you suggesting, then, that the “fix” for peer review is to have the reviewers reproduce the experiments to detect frauds and mistakes? If so, you’ll never find anyone to review a manuscript and, if you did, it would take years to get published. I think everyone agrees that there are flaws to the system, but I’ve yet to hear a method that won’t be equally open to abuse. Maybe open review with identities verified through Facebook accounts is the solution.

  3. Derek Freyberg
    July 3, 2014 at 11:58 AM

    My take on this is not that peer reviewers (referees) should be obliged to reproduce experiments, and I don’t think that’s Sharon’s either.
    However, “extraordinary claims require extraordinary evidence”; restated, papers making extraordinary claims, as the Obokata papers did, should be subject to especially careful review.
    In this case, some of what appeared in the papers was sufficiently odd to people in the field reading them (considering the post-publication kerfuffle) that I don’t see Nature’s claim that it could not have been spotted by peer review as very credible.
    I note that RIKEN’s own report on Obokata’s appeal contains the following:
    “i) When the appellant, after the 2012 Paper submitted to Nature in April 2012 was rejected,
    submitted a paper with similar content (the 2012 Paper with the addition of electrophoresis
    photographs to demonstrate T cell receptor rearrangement [Supplemental Figure 6]) to Science
    in July, she received the following comment from the reviewer: “Moreover, this figure
    has been reconstructed. It is normal practice to insert thin white lines between the lanes taken
    from different gels (lanes 3 and 6 are spliced in). Also I find the leading edge of the GL band
    suspiciously sharp in #2–#5.””
    So a Science reviewer/referee spotted the manipulation of that figure, even if Nature’s referees didn’t.
    I return to my second sentence.
    I think that the journal refereeing process is not well adapted to detecting outright manipulation of data, such as raising synthetic yields in chemistry, though I understand that some journals are putting mechanisms in place to ensure that image data is not obviously manipulated; and crystallographic journals, for example, insist on the authors providing original datasets rather than just results. And, for papers just reporting the latest member of the “methyl, ethyl, propyl, butyl, futile” sequence, I don’t think that by-and-large it should have to be. But when a paper contains conclusions that, if true, upend conventional thinking, I see a certain extra obligation on the journal that publishes it to ensure not merely that the conclusions are supported by the data in the paper as it will appear, but also that the data in the paper as it will appear are, for lack of a better word, true.

    • Jon O
      July 3, 2014 at 12:27 PM

      This is an excellent highlight of the difference between the two celebrity journals of Science and Nature. Of the two, Nature publishes far more of the big frauds.

      In a few years, I fully expect to see software tools for image analysis to detect scientific fraud analogous to the tools now available to detect textual plagiarism. Unfortunately, I feel that peer review will always be susceptible to those seeking to purposefully mislead. The emphasis in cases like this is going to be correcting the scientific record after the results have been made public. And, in most ways, I think this post-review vetting is sufficient to maintain the validity of the scientific record, though it can sometimes lead to grief if people are building upon incorrect studies.

Comments are closed.