SRxA
It's apparently very easy to fool yourself when you're doing stem cell research. And not just yourself—the peer reviewers at leading scientific journals can be easily gulled as well. Just six months ago, Nature published two very high-profile papers by Japanese researcher Haruko Obokata in which she claimed that her team could turn normal mature cells into stem cells by simply bathing them in a mild acid. This would be a huge breakthrough, since stem cells might then be easily manufactured for therapeutic use in individual patients. Alas, the work appears to be too good to be true. The studies have been retracted, according to the Associated Press. From the AP:
On Wednesday, Nature released a statement from Obokata and the other authors of the papers that withdrew the papers. The scientists acknowledged "extensive" errors that meant "we are unable to say without a doubt" that the method works. They noted that studies of the simpler method are still going on by other researchers. Dr. Charles Vacanti of Brigham and Women's Hospital in Boston, another main author, issued a separate statement in which he said he believes the further studies will vindicate the method, which produced what the authors called STAP cells. But another author, Yoshiki Sasai, deputy director of the Riken center, said the errors in the papers meant "it has become increasingly difficult to call the STAP phenomenon even a promising hypothesis." In a statement issued by Riken, he said he was "deeply ashamed" of the problems in the papers. Retractions of papers in major scientific journals like Nature are rare. They can come about because of fraud or the discovery of honest mistakes that undercut the conclusions of research. Publications like Nature routinely have experts review papers submitted by scientists to look for problems. But in an editorial released Wednesday, Nature concluded that its editors and reviewers "could not have detected the fatal faults in this work."
Glory (funding and tenure) tempts researchers to rush results into high-profile journals. Unfortunately, this incentive structure for research seems to have gone considerably awry. For example, in 2012 I reported in my column "Can Most Cancer Research Be Trusted?" that researchers could replicate the results of only six out of 53 landmark cancer research papers.
More researchers and journals should follow the replication system devised by the folks at the Open Science Framework project. Even better, major journals could insist that they will publish "game changing" papers only after the research has been replicated by an outside lab.
Comments