University of Missouri probes possible fraud

Images in Science paper may have been faked.

Nov 14, 2006
Kerry Grens

The University of Missouri has opened an inquiry into whether researchers in the biochemistry department there manipulated images published in a February 2006 article in Science. The data, which were produced by postdocs in R. Michael Roberts' laboratory, challenged the conventional theory that individual blastomeres in early embryos are identical.

The inquiry centers around whether two images in the paper were altered to misrepresent the data, Roberts told The Scientist. "It's been a nightmare," he said. "This is a very difficult and painful time for people in the lab." Roberts said that the paper may ultimately have to be withdrawn.

Last month, Science editor-in-chief Donald Kennedy issued an editorial expression of concern cautioning readers that the results reported in the paper "may not be reliable." "We're anxious to bring this to an end as soon as possible," Robert Hall, associate vice chancellor for research at the University of Missouri, told The Scientist.

The investigation is in the first stage, which consists of an inquiry by three senior university scientists to determine whether fraud has occurred. Hall expects a finding in the next several weeks. If the panel determines that there has been wrongdoing, the investigation will move into the second phase, which determines penalties. The study, conducted by Kaushik Deb, Mayandi Sivaguru and Hwan Yul Yong, who were postdocs in Roberts' lab at the time, examined whether blastomeres from early-stage embryos expressed the transcription factor Cdx2 equally.

"[C]onsiderable debate rages about whether the mature mouse oocyte contains factors localized in such a manner that they direct future cell differentiation," the authors wrote. Their findings stoked that debate. In opposition to the long-held view that blastomeres in early-stage embryos are equivalent, they found Cdx2 expression at the two-cell stage was localized to blastomeres at the vegetal pole of oocytes, and that these differences lead to distinct cell lineages.

Richard Schultz, a biology professor at the University of Pennsylvania, told The Scientist that some researchers studying embryonic development were suspicious early on. "Whenever there's a result that goes against the dogma, it always raises eyebrows, but that doesn't mean the dogma's correct." He said skepticism over the findings escalated when other labs, his own included, failed to replicate the result. Schultz's laboratory wanted to use Cdx2 expression as a marker for investigating the consequences of preimplantation genetic diagnosis, a procedure that removes a blastomere for genetic testing after in vitro fertilization. But he was unsuccessful at reproducing the asymmetrical allocation of Cdx2. "We had difficulty getting it, and dropped the project," he said.

Magdalena Zernicka-Goetz at the University of Cambridge in the UK said the observations in the original study surprised her. "Neither ourselves or other laboratories?have ever seen completely independent lineages for completely different cell types arising at the two-cell stage," she told The Scientist in an email. "Everyone studying the mouse embryo has found that each of the two-cell blastomeres gives rise to both cell types."

Roberts is a member of the National Academy of Sciences and has led research at the University of Missouri for over 20 years. Schultz told The Scientist that Roberts has a "pristine reputation" and that he believes it's highly unlikely that he would have been involved in manipulating the data. "He's an absolutely straight-shooter," he said. "If there was any wrongdoing, it was done without his knowledge." The last time the University of Missouri conducted a formal inquiry and found wrongdoing was in 2000. According to Hall, a postdoc plagiarized an image in a funding proposal to the National Institutes of Health and ultimately resigned.