Fraud case we might have seen coming

Eugenie Samuel Reich

Nobody likes to hear 'I told you so', not least over something that has had far-reaching consequences. But when David Baker read an article reporting that Luk Van Parijs, a former associate professor of immunology at the Massachusetts Institute of Technology (MIT) in Cambridge, had been convicted of grant fraud, he felt compelled to leave a brief anonymous comment noting that he had raised concerns over Van Parijs' data 14 years earlier.

Baker, a neuroimmunologist at Queen Mary, University of London, who voiced his concerns in an e-mail to the Journal of Experimental Medicine (JEM) in 1997 about a paper1 authored by Van Parijs, feels that had the journal acted — perhaps by contacting other authors on the paper or by referring the matter to Harvard Medical School in Boston, Massachusetts, where Van Parijs was then a graduate student — Van Parijs might have modified his behaviour, or perhaps gone on to a career outside science. That, in turn, might have prevented the later impact of his actions on the MIT trainees who, in 2004, brought allegations against their mentor, and on Van Parijs himself, who was sentenced in June to six months home detention.

"It's sad. If he'd been called to task earlier, he wouldn't be paying the price today," says Baker. This, he notes, shows the importance of nipping scientific fraud in the bud.

Figures ignored
MIT fired Van Parijs in 2005, after a year-long investigation of the concerns raised by members of his lab. It quickly emerged that several publications dating back to 1997 contained partly duplicated and manipulated data — later declared to be misconduct by panels at two institutions where Van Parijs had formerly worked and studied.

A common assumption in cases of such blatant duplication is that other scientists fail to notice the problem after publication because they tend to focus on the text of a scientific paper, only briefly checking the figures to see whether the data support them. Often, it is only when researchers have problems reproducing work that they look at the data more critically, says John Dahlberg, head of the Division of Investigative Oversight at the US Office of Research Integrity. However, Baker's experience highlights an additional factor: even when scientists do spot and report problems, their efforts to intervene are not always successful.

Van Parijs' inconsistent data were brought to Baker's attention by one of his graduate students, who noticed that a series of flow cytometry plots purportedly from different samples appeared to have been partly duplicated.

When Baker looked at the figure, he realized within seconds that its components had been manipulated, and e-mailed editors at the JEM. His e-mail, provided in copy form to Nature, describes problems in the data in detail, and ends by saying: "This strikes me as odd, maybe you should look at the figure?"

Serial offender
Baker did not receive a response from the JEM. Over the following few years, his students occasionally checked to see what Van Parijs had published recently, and whether anything looked as if it had been manipulated. In the late 1990s, members of Baker's lab noticed manipulated data in three further papers published by Van Parijs2,3,4. All three were later retracted.

Despite the concerns in his lab, Baker did not mention the issue to colleagues elsewhere, partly because Van Parijs was not in his subfield of neuroimmunology, so the matter was not scientifically relevant, and also because he'd tried to intervene already, without success. "I felt I'd done my bit," he says. Eventually, he stopped tracking Van Parijs, but in 2009 looked him up again and saw that he had been censured by the US government for misconduct, and that his the JEM paper had been retracted. Baker admits that, when he read about Van Parijs' trainees, he regretted not doing more.

Baker suspects others may also have noticed the figure and not spoken up. "People are afraid to stick their head above the parapet because the whistleblower never comes off well," he says.

In 2009, Baker wrote to the JEM again, and received a reply from senior editor Heather Van Epps, who joined in 2004, apologizing for the journal's lack of response to his 1997 e-mail. Van Epps told Nature that the JEM records do not stretch back far enough for her to establish what happened to Baker's initial inquiry. She declined to comment on Baker's suggestion that action by the JEM could have averted the later, major fraud scandal. She also notes that, in 1997, the JEM did not yet have professional editors. Two scientists who served as editors at the time told Nature that they did not recall reading Baker's e-mail, which was sent to a communal e-mail address, and were unsure how such an e-mail might have been handled.

Mike Rossner, the executive director of the New York-based Rockefeller University Press, which publishes the JEM, was out of town and did not respond to a request for comment. Since 2002, Rossner has led a high-profile effort to screen for manipulated images in papers before publication.

To this day, not all of the papers published by Van Parijs containing suspicious duplications have been corrected or retracted. One5, published in Immunity in 2003, contains what looks to be a partly duplicated, manipulated image, first published as part of an entirely different experiment in the Journal of Immunology in 20026, and found to be falsified by the Office of Research Integrity. Two inquiries by Nature to editors at Immunity noting the problem and asking about the fate of the paper have not been answered.

References

1.Van Parijs, L. et al. J. Exp. Med. 186, 1119-1128 (1997). | Article | PubMed | ChemPort |
2.Van Parijs, L., Peterson, D. A. & Abbas, A. K. Immunity 8, 254-274 (1998). | Article |
3.Van Parijs, L. et al. Immunity 11, 281-288 (1999). | Article | PubMed | ISI | ChemPort |
4.Van Parijs, L., Refaeli, Y., Abbas, A. K. & Baltimore, D. Immunity 11, 763-770 (1999). | Article | PubMed | ISI | ChemPort |
5.Layer, K. et al. Immunity 19, 243-255 (2003). | Article | PubMed | ISI | ChemPort |
6.Kelly, E., Won, A., Refaeli, Y. & Van Parijs, L. J. Immunol. 168, 597-603 (2002). | PubMed | ISI | ChemPort |

nature.com