Review of Cloning Paper Prompts Questions Nearly 8 years after a massive fraud perpetrated by South Korean stem cell researchers, the embryo cloning breakthrough they claimed was back in the news this month. A different group published a paper reporting that they had cloned human embryos and derived stem cells from them. But incredibly, the new paper has come under scrutiny itself for duplicated images. So far there’s no indication that the recent work, by a well-respected Oregon-based team that studies primate cloning, is fraudulent or even wrong. “It was just an honest mistake,” says the study’s senior author, Shoukhrat Mitalipov of the Oregon National Primate Research Center in Beaverton, speaking of a pair of image mix-ups. The first author, lab member Masahito Tachibana, is “devastated,” Mitalipov says. But he and his colleagues are confident that the work will stand up to scrutiny, he says. The cell lines that his team derived are being distributed to other labs, which he expects will soon confirm the results. Mitalipov’s paper, published online on 15 May by Cell, used cloning to create personalized human embryonic stem (ES) cells from human skin cells (Science, 17 May, p. 795). The errors—several images in the paper appeared more than once under different labels—are another black mark on how high-profile papers are vetted, or not, before being published. They raise troubling questions that have dogged scientific publishing for years. For instance, how much responsibility do journals and reviewers bear for detecting problematic images in papers, whether honest errors or not? How is it possible that the very same result that unraveled so spectacularly in 2005 was not exhaustively reviewed when it was submitted for publication a second time?

1026

Cell issued a statement the day after an anonymous commentor posted about the problem. The journal admitted to “some minor errors made by the authors.” And it defended its unusually quick review; the research was formally submitted on 30 April and accepted on 3 May. In an e-mail to Science, the editor-in-chief of Cell, Emilie Marcus, writes that “the reviewers returned their comments quite rapidly and given their Double trouble. Several images were used twice, and one set (below) was mislabeled in a paper describing human cloned embryos by Shoukhrat Mitalipov (left) and his colleagues.

recommendations for acceptance, with our focus on author service and because of the interest and importance of the work, we then decided to take special efforts to keep the delay between acceptance and publication to a minimum to reduce the likelihood that the findings would come out first in the popular press.” That desire to snag the hottest of hot papers, and bask in their limelight, is often at odds with what it takes to fact-check them as thoroughly as possible. “These things happen,” says Robin Lovell-Badge, a stem cell researcher at the MRC National Institute for Medical Research in London, though he notes that time pressure raises the chance that mistakes will slip through. “I think the fault lies with the combination of the

31 MAY 2013

VOL 340

SCIENCE

Published by AAAS

authors, the editors, and the reviewers. It’s been missed by three sets of people,” LovellBadge says. Inconsistencies in the Mitalipov paper first surfaced on 22 May on PubPeer, a site for postpublication peer review where commenters can describe concerns about a paper and authors can respond. Although it’s hardly the first time that anonymous voices on the Internet have proven critical, “I don’t think we should be relying on the crowd to be policing” research papers, says Hany Farid, a computer scientist at Dartmouth College who is working to commercialize software that can detect duplicated images and other manipulations. Reviewers are already “overburdened,” and Farid believes it’s the journals that must take this on. At Science, which in 2004 and 2005 published the stem cell work by South Korean scientist Woo-Suk Hwang that turned out to be fraudulent, Executive Editor Monica Bradford says that she empathizes with her counterparts at Cell. “I know what position they’re in. You go from this amazing high … to ‘Oh my God.’ … You feel like, ‘How did we make these mistakes?’ ” Bradford was at Science during the Hwang debacle and was interviewed at the time by the authors of this article about the journal’s response. The Hwang fraud, she says now, had a lasting impact. Science commissioned an external report led by Stanford University chemist John Brauman to examine what it might do differently in the future. “It is essential to develop a process by which papers that have the likelihood of attracting attention are examined particularly closely for errors, misrepresentation, deception, or outright fraud,” wrote the authors of the Brauman report, as it was called. After the Hwang case, which involved deliberate image manipulations, Science began checking all figures in papers at the revision stage and launched previously planned routine analysis of images using Photoshop software. “My level of trust is way different than when I started this job,” Bradford says. “In the area of stem cells we’ve been very cautious, maybe too cautious.” But other papers that later were dis-

www.sciencemag.org

Downloaded from www.sciencemag.org on January 9, 2014

BIOMEDICINE

CREDITS (TOP TO BOTTOM): COURTESY OF OREGON HEALTH & SCIENCE UNIVERSITY; M. TACHIBANA ET AL. CELL (ADVANCED ONLINE EDITION)

NEWS & ANALYSIS

NEWS credited have continued to slip through, such as work claiming that a virus called XMRV causes chronic fatigue syndrome. It was later retracted. “It’s like whack-a-mole. We keep trying,” Bradford says, “but we are far from perfect, so that’s why I really have sympathy for Cell.” The software that Science and some other journals use isn’t designed to detect duplicated images, although it can make them easier to spot. Rather, it reveals various inappropriate modifications made to an image to

enhance it. Marcus declined to say whether Cell uses image analysis technology. In 2006, she expressed reluctance to cast blanket suspicions in a story in The New York Times about screening images. “Why say, ‘We trust you, but not in this one domain?’ And I don’t favor saying, ‘We don’t trust you in any,’ ” Marcus told the Times reporter. Stem cell scientists expressed dismay that their field is again under scrutiny. Some say that they are working hard to replicate Mitalipov’s results. Even if the image dupli-

cations and mislabelings are quickly corrected—as Cell and the authors say they’re working to do—Farid warns that cases like this one can ripple well beyond the labs parsing the cells. “Suddenly the public starts wondering what the hell we’re doing,” he says. It’s important and natural for science to move slowly, he says, perhaps especially so when the results seem remarkable. After all, Farid says, “This isn’t cable news.” –JENNIFER COUZIN-FRANKEL AND GRETCHEN VOGEL

U . S . I M M I G R AT I O N P O L I C Y

CREDIT: WIN MCNAMEE/GETTY IMAGES

Visa Reform Advances in Senate as House Offers STEM Ideas Efforts to overhaul U.S. immigration policy numerous obstacles, however. Some labor from 65,000 to 110,000—were needed to often generate more show than substance. specialists, for example, worry that the Sen- win his support in committee. But he said Last week, they produced both: a three-ring ate bill does too little to protect U.S. STEM that he will take a fresh look when the meacircus of activity that pushed reform ahead workers from immigrant competition. Provi- sure comes before the full Senate. without deciding what Congress ultimately sions allowing more high-skill guest workThe deal that Hatch struck with Senawill do. ers will “depress wages in STEM fields and tor Charles Schumer (D–NY) displeased Most attention focused on the U.S. Sen- put U.S. workers at a disadvantage,” predicts groups representing U.S. engineers and ate, where a committee last week won economist Hal Salzman of Rutgers Univer- other skilled workers, who would prefer bipartisan agreement on a path to citizen- sity in New Brunswick, New Jersey. fewer H-1Bs. But STEM advocates welship for some 11 million undocumented Such complaints got limited traction as comed another Hatch amendment adopted immigrants. By a 13 to 5 vote, the panel on the Senate Judiciary Committee spent 5 days by the panel that steers more visa fees— 21 May approved a bill that also makes it considering hundreds of proposed changes potentially tens of millions of dollars easier for foreign-born scientists and engi- to a bipartisan proposal released last month annually—to STEM programs funded by neers to live and work in the United States (Science, 26 April, p. 415). In one move, the the U.S. Department of Education and run and taps visa fees to create new funding panel weakened rules designed to encourage by states. They were also glad that senastreams for STEM (science, technology, U.S. high-tech companies to hire U.S. work- tors didn’t tinker with provisions that would engineering, and mathematics) education ers before turning to foreign guest workers automatically give a green card to a foreign and training programs aimed at providing on temporary visas known as H-1Bs. Sena- student who earns an advanced degree from more domestic workers with STEM skills. tor Orrin Hatch (R–UT), siding with high- a U.S. university in all STEM fields. But members of the U.S. House of Repre- tech companies in his state, said that the A narrower high-skill immigration bill sentatives also shared the limelight. Repub- changes in hiring temporary workers—the introduced on 24 May by senior House lican leaders demonstrated their piecemeal bill would raise the annual H-1B ceiling Republicans includes similar provisions. approach to immigration reform Two differences are that it would with a bill that would ease the allow more H-1Bs—155,000— entry of skilled foreign labor and cap the number of green in ways that paralleled the Sencards for foreign-born STEM ate legislation. And a bipartisan graduates at 55,000. group of House members continSuch a standalone Repubued work on a broader immigralican bill has little chance of tion proposal that is expected to winning Senate approval, howinclude STEM provisions. ever. But it could help a biparti“We’re seeing a convergence san House group shape a more on some policies that try to … comprehensive immigration bill retain skilled STEM immigrants that might fly. But that draft is and improve STEM education,” still under wraps, and months of says Scott Corley, executive negotiations have yielded only director of Compete America, broad principles. The group says a coalition of business and acathat it might have a bill ready demic groups that backs many of next month, when the Senate is Temporary fix. Senator Orrin Hatch (R–UT) (center) won backing for weaker the proposals. expected to act. rules on temporary visas to high-skill foreign workers. The overhaul effort faces –DAVID MALAKOFF www.sciencemag.org

SCIENCE

VOL 340

Published by AAAS

31 MAY 2013

1027