The stem cell research world was set on its ear 29 January when Japanese and American scientists reported an astoundingly simple way to generate stem cells that can theoretically develop into all of a body’s cells. But anonymous bloggers and some experts have spotted possible image manipulation in the papers. The lead researcher’s institution is investigating.
Haruko Obokata of the RIKEN Center for Developmental Biology (CDB) in Kobe, Japan, and colleagues at other Japanese institutions and at Harvard Medical School in Boston reported that simply subjecting blood cells from newborn mice to a moderately acidic environment for 25 minutes and then tweaking culture conditions could generate pluripotent stem cells capable of developing into nearly all of a body’s cell types. They described what they dubbed stimulus-triggered acquisition of pluripotency (STAP) cells in a research article  and a research letter  published online in Nature on 29 January.
Just a week later, anonymous bloggers at the PubPeer website, a forum for postpublication discussion of scientific results, started pointing out  anomalies in image 1i in the research article. To some eyes, it appears that the center lane of a blot image has been spliced in. On 13 February, PubPeer contributors zeroed in on the research letter, claiming that part of figure 1b, showing the placenta of a STAP chimera, may have been rotated and reused  in figure 2g.
In e-mails responding to questions from ScienceInsider, the public relations office at RIKEN's headquarters in Wako, near Tokyo, wrote that they were aware of the website but launched their investigation on 13 February in response to a notification received from an outside expert. The investigative team includes both RIKEN and outside scientists. RIKEN could not say when the investigation would be completed, who leads it, or if it focuses only on the images flagged on PubPeer.
A public relations official at CDB said Obokata and her colleagues were directing all queries to RIKEN headquarters.
On 13 February, a Japanese blog pointed to possible image manipulation  in a 2011 paper in Tissue Engineering by Obokata and others. On the other hand, on 14 February, a Japanese blogger posted a note  stating that "the problems of some images of the research article could be the result of image compression noise."
The RIKEN public relations officials wrote that their investigation covers only the two Nature papers.
Those two papers were groundbreaking because they put forward a method for generating stem cells far simpler than any previously reported, a development that could advance regenerative medicine, in which scientists try to grow replacement tissues as a treatment for diseases and injuries.
The image problems might not mean the reported results are invalid.
"At the moment, it is thought that the research results published in Nature are sound," the public relations office wrote.
Other labs are already trying to replicate the work, and some have posted their experiences  on a stem cell blog. None has yet said they have successfully reprogrammed cells, but most have not used the same cells Obokata used.
*Update, 17 February, 3 p.m.: A sentence referring to a second blog post in Japan discussing possible compression of the images was revised as a result of a clarification of its translation into English.