Ethics Scandal at Duke
Duke University has been rocked by an ethics scandal in its biochemistry department, and I learned about it by reading an Indian blog on higher education, of all places (thanks to Abi and his nanopolitan blog for the coverage of this sad affair - read his post dated May 22). The full story, published in Nature, is available for free download in pdf format on the writer's website; I summarize it below. A graduate student, Mary Dwyer, decided to join the lab of Dr. Homme Hellinga at Duke when she was looking for a doctoral adviser, although she was warned against him; their collaboration led to several publications in the prestigious scientific journal Science. (Hellinga studies how amino acids encode function in proteins.) Dwyer graduated in 2004 and moved on to a post-doctoral position in another lab at Duke in 2005. A few years later, a researcher contacted Hellinga and announced he could not duplicate the results. Hellinga then proceeded to accuse Dwyer of scientific misconduct in the Fall of 2007 - in other words, he accused her of falsifying the data in her experiments. Duke, at his request, investigated her. The woman still had the notebooks where she had recorded the data and, while she made mistakes in her experiments (the products she was testing were contaminated; hence, it seemed the enzymes showed activity although they did not), was cleared of scientific misconduct.
Dwyer now says that she had doubts that the project - about redesigning proteins to make them active enzymes - would pan out and warned Hellinga that there was too much variability in the results, but Hellinga rushed her to publish. There are, of course, two sides to every story. It is in Dwyer's interest, if she has any hope of getting a tenure-track position after this debacle - which saw Hellinga withdraw her articles from Science with much publicity - to portray herself as a victim. The Nature article is heavily biased in favor of Dwyer. For the case at hand, it is interesting to note that many students in the biochemistry department at Duke, commenting anonymously in the blogosphere, have confirmed that Hellinga is a first-rate jerk, and Dwyer herself had been advised by students in the lab not to join. Among Hellinga's first twelve graduate students, only two received their PhD. (It is not clear whether the others changed labs or quit graduate school.) While attrition exists in any doctoral program, those statistics suggest Hellinga's advising was lacking in quality. In the end, Dwyer probably made serious but honest mistakes, then let the articles become published despite her concerns. After all, their publication was in her interest, not only to graduate but to get a good position in another lab; she might, however, have hoped they were true somehow, since other researchers were bound to (attempt to) duplicate such high-profile results sooner or later. It doesn't make sense for someone who wants to stay in academia to let papers with dubious findings become published. Dwyer is famous all over the blogosphere now, and not in a good way.
Her most important mistake happened years before she performed the experiments, when she disregarded warnings and decided to work in Hellinga's lab. A lot of graduate students, faced with the prospect of working with a famous but difficult adviser, will convince themselves that other students deserved to get in trouble somehow, and none of this will happen to them. They work hard, they mean well, who wouldn't want to work with them? The truth is, the relationship between advisee and adviser depends not only on the student's motivation but on the professor's character. A professor will not work well with every student. While it's hard for a twenty-two-year-old graduate student to realize this, one would expect a competent, seasoned adviser to stay away from students he feels won't be a good match with him, rather than trying to profit from their intellectual skills despite the mismatch. Hellinga in particular comes across as someone any student in their right mind would run away from as fast as they can. I don't think you can put the blame regarding the experimental mistake solely on him - Dwyer was a senior graduate student by the time she did the experiments, he trusted her to tell him what was wrong - but his behavior once the mistake was uncovered is a disgrace. Dwyer should have been given the benefit of the doubt until Duke had completed its investigation.
Contrast Hellinga's behavior with that of Dalibor Sames, a Columbia University professor who, in 2006, retracted six papers (two in March and four in June), which had been co-authored with a former doctoral student, Bengu Sezen, who, like Mary Dwyer, had left the lab by the time the allegations were made. (However, the six papers co-authored by Ms Sezen span many years.) The New York Times coverage mentions that "The retractions came after the experimental findings of the papers could not be reproduced by other researchers in the same laboratory" and notes that "Columbia has opened an inquiry into why the experiments were not reproducible." The adviser, Dalibor Sames, makes sensible points on the difficulty (embarrassment) of retracting a paper but stays above the fray, leaving it to Columbia to decide whether Ms Sezen acted deliberately. His focus appears to be on science rather than finger-pointing. Unsurprisingly, the former graduate student denies any wrong-doing, explaining that she had not been contacted by the university and knew nothing about the case before the reporter got in touch with her. The reporter notes, however, that "although the reporter's query did not list the three articles in question, Dr. Sezen noted them in her reply" and "a university spokeswoman said Dr. Sames had not received any recent communication from Dr. Sezen, and added, 'She has not acknowledged our recent attempts to reach her.'" After graduating from Columbia with a degree in chemistry, Ms Sezen enrolled in the PhD program in molecular biology at the University of Germany in Heidelberg. Columbia's inquiry appears to have been inconclusive, since there is no information about its findings on the Internet.
This all pales in comparison with the clearly unethical behavior of several other researchers (and yes, already doctors when they committed fraud), described in a January 2007 article in Nature. The cases are so egregious it is hard to single out one or two of them. Jon Sudbo and Jan Hendrik Schoen, for instance, falsified data on a grand scale (the former creating nine hundred fictitious patients in a case study and fabricating data in over thirty papers, the other falsifying data in "only" sixteen papers). I find it hard to understand why anyone would behave like that and not expect to get caught, although it seems it took the scientific community a while to catch on. Unfortunately, there is little incentive for other teams to replicate experiments once the discovery has been announced - they won't be able to take any of the credit. But maybe, in the same way as the National Science Foundation now requires each grantee to include the cost of traveling to the grantees' conference in his or her budget, researchers in experimental science will one day be forced to provide an outside team with the agents and protocol needed to replicate the experiments before results can be published.
Dwyer now says that she had doubts that the project - about redesigning proteins to make them active enzymes - would pan out and warned Hellinga that there was too much variability in the results, but Hellinga rushed her to publish. There are, of course, two sides to every story. It is in Dwyer's interest, if she has any hope of getting a tenure-track position after this debacle - which saw Hellinga withdraw her articles from Science with much publicity - to portray herself as a victim. The Nature article is heavily biased in favor of Dwyer. For the case at hand, it is interesting to note that many students in the biochemistry department at Duke, commenting anonymously in the blogosphere, have confirmed that Hellinga is a first-rate jerk, and Dwyer herself had been advised by students in the lab not to join. Among Hellinga's first twelve graduate students, only two received their PhD. (It is not clear whether the others changed labs or quit graduate school.) While attrition exists in any doctoral program, those statistics suggest Hellinga's advising was lacking in quality. In the end, Dwyer probably made serious but honest mistakes, then let the articles become published despite her concerns. After all, their publication was in her interest, not only to graduate but to get a good position in another lab; she might, however, have hoped they were true somehow, since other researchers were bound to (attempt to) duplicate such high-profile results sooner or later. It doesn't make sense for someone who wants to stay in academia to let papers with dubious findings become published. Dwyer is famous all over the blogosphere now, and not in a good way.
Her most important mistake happened years before she performed the experiments, when she disregarded warnings and decided to work in Hellinga's lab. A lot of graduate students, faced with the prospect of working with a famous but difficult adviser, will convince themselves that other students deserved to get in trouble somehow, and none of this will happen to them. They work hard, they mean well, who wouldn't want to work with them? The truth is, the relationship between advisee and adviser depends not only on the student's motivation but on the professor's character. A professor will not work well with every student. While it's hard for a twenty-two-year-old graduate student to realize this, one would expect a competent, seasoned adviser to stay away from students he feels won't be a good match with him, rather than trying to profit from their intellectual skills despite the mismatch. Hellinga in particular comes across as someone any student in their right mind would run away from as fast as they can. I don't think you can put the blame regarding the experimental mistake solely on him - Dwyer was a senior graduate student by the time she did the experiments, he trusted her to tell him what was wrong - but his behavior once the mistake was uncovered is a disgrace. Dwyer should have been given the benefit of the doubt until Duke had completed its investigation.
Contrast Hellinga's behavior with that of Dalibor Sames, a Columbia University professor who, in 2006, retracted six papers (two in March and four in June), which had been co-authored with a former doctoral student, Bengu Sezen, who, like Mary Dwyer, had left the lab by the time the allegations were made. (However, the six papers co-authored by Ms Sezen span many years.) The New York Times coverage mentions that "The retractions came after the experimental findings of the papers could not be reproduced by other researchers in the same laboratory" and notes that "Columbia has opened an inquiry into why the experiments were not reproducible." The adviser, Dalibor Sames, makes sensible points on the difficulty (embarrassment) of retracting a paper but stays above the fray, leaving it to Columbia to decide whether Ms Sezen acted deliberately. His focus appears to be on science rather than finger-pointing. Unsurprisingly, the former graduate student denies any wrong-doing, explaining that she had not been contacted by the university and knew nothing about the case before the reporter got in touch with her. The reporter notes, however, that "although the reporter's query did not list the three articles in question, Dr. Sezen noted them in her reply" and "a university spokeswoman said Dr. Sames had not received any recent communication from Dr. Sezen, and added, 'She has not acknowledged our recent attempts to reach her.'" After graduating from Columbia with a degree in chemistry, Ms Sezen enrolled in the PhD program in molecular biology at the University of Germany in Heidelberg. Columbia's inquiry appears to have been inconclusive, since there is no information about its findings on the Internet.
This all pales in comparison with the clearly unethical behavior of several other researchers (and yes, already doctors when they committed fraud), described in a January 2007 article in Nature. The cases are so egregious it is hard to single out one or two of them. Jon Sudbo and Jan Hendrik Schoen, for instance, falsified data on a grand scale (the former creating nine hundred fictitious patients in a case study and fabricating data in over thirty papers, the other falsifying data in "only" sixteen papers). I find it hard to understand why anyone would behave like that and not expect to get caught, although it seems it took the scientific community a while to catch on. Unfortunately, there is little incentive for other teams to replicate experiments once the discovery has been announced - they won't be able to take any of the credit. But maybe, in the same way as the National Science Foundation now requires each grantee to include the cost of traveling to the grantees' conference in his or her budget, researchers in experimental science will one day be forced to provide an outside team with the agents and protocol needed to replicate the experiments before results can be published.
Comments