Mislabeled Images Bedevil Landmark Cloning Paper

On defense. Shoukhrat Mitalipov says mistakes in cloning paper don't invalidate the results

Oregon Health & Science University.

One day after a prominent paper in the journal Cell was flagged for image duplication, the main author and the journal say that the problems arose from simple mislabeling of images and do not invalidate the results. They also defended the unusually rapid review of the paper, which was accepted only 4 days after official submission and published online 12 days later.

The work, led by Shoukhrat Mitalipov of the Oregon National Primate Research Center in Beaverton, was notable for two reasons: It's the first time anyone has used cloning to create personalized human embryonic stem (ES) cells, and it's the same result that was described back in 2004 and 2005 by a group of South Korean scientists in what turned out to be one of the world's most notorious cases of scientific fraud.

Stem cell scientists were initially delighted, with one telling Science that the work was a "hard-won triumph after many years of diligent research." After a commenter posted on a site called PubPeer, alleging duplicated and mislabeled images in the paper, delight turned to dismay. "It's a shame that this important area of research has come under scrutiny once again," writes Kevin Eggan of Harvard University to ScienceInsider in an e-mail.

Mitalipov spoke with ScienceInsider this afternoon, and says that he and the other co-authors simply overlooked the mistakes, in part because the images in question are meant to show that the cells are similar. "With the naked eye, it's very hard to see if this is the same image or a different image." He says that he is curious whether the PubPeer contributor used image-checking software to catch the duplications. "I wish we had that software to run the paper through," he says.

The first set of duplications involves three figures. The images are meant to show that the ES cells from cloned embryos look similar to those derived from IVF embryos, suggesting the cloned ES cells are the real thing. But the same image seems to show up twice under a different label: once as an an ES cell line derived from a cloned embryo, and elsewhere in the paper as a control ES cell line from an IVF embryo. Another image in that same set, figure 6, appears as a cloned cell line but also shows up in a supplemental figure labeled as the control line. Mitalipov, who spoke with ScienceInsider this afternoon, says that first author Masahito Tachibana deliberately used the images twice, but accidentally reversed the labels in figure 6.

The second duplication appears in supplemental figure S6, where a scatterplot purporting to show gene expression similarities between the cell lines was used twice. Mitalipov says that the wrong scatterplot was used, and the correct one will be published in an erratum. The original microarray data are publicly available, he notes.

Mitalipov argues that the key data proving whether the cell lines are really derived from cloned embryos is unaffected by these errors. What's most important is whether the ES cells' mitochondrial DNA matches that of the egg cell donor and whether the nuclear DNA matches the cell that was cloned. The researchers deliberately chose a widely available cell line for their experiments, Mitalipov says, making it easy for outside labs to try to confirm the results. He says that he is ready to ship the cloned ES cell lines to several labs that have requested them as soon as the Oregon institutional review board signs off on the transfers, which could happen in a matter of days. (Federal funding restrictions forbid National Institutes of Health-funded labs from working on the cells, because they were derived via cloning, so recipients also need to show that they have a legal place to work with the cell lines.)

Cell , meanwhile, sought to defend itself. Editors from the journal declined to speak with Science, but spokesperson Mary Beth O'Leary released a statement, noting that "it seems there were some minor errors made by the authors. … We do not believe these errors impact the scientific findings of the paper in any way."

Several stem cell experts tell ScienceInsider that the images in question aren't key to the paper's conclusions. Nevertheless, Eggan writes, "we'll likely have to wait until the cell lines in question are validated by others, or an independent group replicates the [cloned ES cell] finding before we'll know for sure." Dieter Egli of the New York Stem Cell Foundation in New York City tells ScienceInsider that he and his colleagues are already working to replicate Mitalipov's claims.

Cell was not able to address one key question before our deadline: What was the rush in securing the paper? It had an unusually rapid turnaround, submitted on 30 April and accepted on 3 May. "The reviewers graciously agreed to prioritize attention to reviewing this paper in a timely way," Cell's statement read. "It is a misrepresentation to equate slow peer review with thoroughness or rigor, or to use timely peer review as a justification for sloppiness in manuscript preparation."

In a story by Nature, Mitalipov seemed to imply that he pressed for rapid review because he wanted to present the work at a conference. But that meeting isn't until mid-June, and Cell has previously permitted authors of high-profile papers to publish in the journal after describing their research at a meeting.

Mitalipov tells ScienceInsider that although he mentioned the conference when he submitted the paper, it wasn't the reason for speedy review, nor was competition from another group. He says he first sent the paper to the journal "5 or 6 days" before the official 30 April submission date as part of a presubmission inquiry. After Cell editors indicated that they were interested, he says, he officially submitted via the website on 30 April. He says the journal simply asked reviewers to look at the paper promptly, "and they did it in a day." The reviewers had only minor critiques, he says, which the authors were able to address quickly. Cell may have been worried about news of the paper leaking, he says, which might have prompted the publication online shortly after acceptance. The paper is scheduled for publication in the 6 June print issue of the journal.

Tachibana is "devastated" by the mistakes, Mitalipov says. Still, the senior scientist is confident that their results will soon be confirmed. "We have the cell lines. We can show what the mitochondrial [DNA] data is, and what the nuclear data is," Mitalipov says. He and his co-authors are now combing through "every dot" in the paper to make sure that there are not any more undiscovered errors before they submit an official correction.