Monday, September 10, 2007

The Kundu Controversy: Research Ethics and Regulation

Divya Gandhi's op-ed in The Hindu last week brought to wider public focus the on-going controversy regarding manipulation in a research paper that has roiled the Indian scientific community for some time now. As pointed out therein, much has been written on it across the blogosphere and in letters to Current Science, an Indian scientific journal run in association with the Indian Academy of Sciences.


Two papers were published in the reputed Journal of Biological Chemistry (JBC) by a doctoral student Hema Rangaswami with her adviser, Gopal C. Kundu from the National Center for Cell Science (NCCS), Pune, the first in 2004 and the second in 2005. The journal, upon receipt of an anonymous e-mail sent allegedly by a former student of Kundu charging that some figures in the second paper were not from new experiments but actually plagiarized from their own previous publication, investigated the complaint and subsequently withdrew the second paper informing the authors that it contained "data that was reproduced without citation and with different labeling", an ethical violation that constituted 'deliberate misrepresentation'. The Society for Scientific Values (SSV), an organization of current and former scientists that claims the upholding of ethics in science as its primary mission and has played a key role in this issue, adds several additional details that have been quoted by other news portals as well: that soon after receipt of the allegations, an internal committee in the institute held the authors guilty and sought withdrawal of the papers and secondly, that Kundu himself, wrote to the journal initially suggesting withdrawal but later sent a second letter alleging that his previous communication was done under duress and insisting that he stood by his publication and it not be rescinded. Kundu's protestations led the Director to ask for an external committee which was then constituted by the Department of Biotechnology (DBT) and comprised of seven members headed by the former Director of the Indian Institute of Science (IISc), G. Padmanabhan. This committee concluded in a brief report (posted on the SSV website), after examining evidence presented by Ms. Rangaswami, that the allegations had no basis and that the e-mails had been sent with the 'malicious intent to spoil the reputation of NCCS'. These conclusions being contrary to the findings of the JBC and the SSV's own analyses (which have both refused to modify their own positions), the DBT demanded a second report from the committee which then compiled a more detailed one running to 118 pages upholding its previous findings. Gandhi's article states that the DBT plans to submit it to a three-member committee for another review.


Before going into the specifics, for those unfamiliar with the field, a few technical details of the science itself are in order. Protein analyses in laboratories are often done by 'blotting', a process that can be used to separate different proteins as well as to identify them individually based on their properties when they appear on the 'blot' as 'bands'. This protein profile on the blot is captured on X-ray film and constitutes the actual evidence used for analysis, also known as 'raw data'. For the purpose of publication, the film itself is photographed and the picture so obtained, edited using digital image processing software to improve clarity. Photo editing is now-a-days universally employed - few pictures that get published are un-edited and the extent of editing depends on technical factors such as visibility of the 'bands', background exposure of the film, etc. At some point, however, a line is crossed and editing can actually start to affect the evidence sought to be presented or in other words, misrepresent it. This is what journal editors are concerned about. JBC, in an editorial shortly after this incident, highlighted this problem even pointing out several types of fraud including the one alleged here that had been detected.


The full picture is not available as neither the report of the JBC's internal investigation nor the detailed report of the Padmanabhan committee appear to have been made public. Conflicting conclusions coupled with this lack of transparency appear to have fueled this controversy. Rahul Siddharthan writing in Current Science showed two blots from the two papers (which do appear identical) and suggested that it is quite implausible for two separate experiments to ever yield such identical pictures. Padmanabhan, in his response in the same issue, presented un-edited pictures of the original X-ray films handed over to the committee and pointed out the differences between the two (which are equally obvious to me). The question then, as Mr. Siddharthan raised, is whether the pictures in the paper were actually taken from the X-ray films purported to represent the original data. Padmanabhan dismissed this query stating simply that he considers such fraud improbable for no conceivable motive exists for such elaborate deception and also argues that many laboratories, running such blots routinely using the same apparatus over and over again, could quite possibly end up with pictures that appear very similar to the naked eye. Veracity of the 'raw data' has also been challenged. Apparently, the internal committee that originally looked into the matter was told that laboratory notes, which are expected to be meticulously maintained as a routine matter, were not available and yet, 'raw data' was provided by Ms. Rangaswami after her arrival. This has led to doubts whether the original data submitted to the Padmanabhan committee was indeed genuine. Padmanabhan on the other hand, blamed the internal committee for not being "a little more mature, giving adequate time and opportunity to all concerned to defend with adequate proof."


The reason for disparate conclusions appears to be the different methods and evidence being examined by different bodies. The first step to resolve this must be to make public all the details of the various investigations done so far. I agree with SSV that truth must ideally be determined from the published evidence itself. Even if the authors' data and conclusions are all fully substantiated and evidence of falsification of 'raw data' does not exist, no journal, being concerned foremost with its own reputation, can be expected to condone or overlook the practice or even to take a lenient view of the matter in light of other mitigating facts. Whether the committee has actually followed this modus operandi is unclear at this time but Padmanabhan's arguments made in his letter to Current Science based on motivation and likelihood are beside the point. However, he might well be right when he says that the truth of the allegation cannot be unequivocally determined from the figures alone. Pictures of protein blots are extensively potrayed in the literature and a determination of plagiarism or photo-fraud beyond reasonable doubt might well be impossible to show with the investigative methods available currently. That raises a significant point: what is the standard of evidence that must be demanded in deciding the issue? Should institutional committees set their own standard or simply take the cue from journal editorial boards? As a general rule, withdrawal of a publication is a serious blot on a scientist's integrity and usually ends up with expulsion from the institution. It also spells the end of his/her academic career as few others are willing to hire the person and journals will be wary of accepting his/her submissions again. Institutions are expected to act promptly and follow through with such action to safeguard their own reputation. An article in Nature on this matter posted a picture of NCCS and the caption below read, "Reputation on the line?" Such pressures usually compel them to play it safe and force the person out raising the difficult question whether institutional disciplinary boards enjoy much leeway in laying down their own rules. Seen from that perspective, the Padmanabhan committee's decision to stand up for its belief and defend the scientist is laudable but I wonder how much good that can do in the long term for either the institute or the scientists themselves.


This case must prompt editorial boards to ask how best to check such a practice. Demanding submission of all relevant original data might be a start. Calls for a formal mechanism to handle allegations of malpractice have arisen. The multitude of committees and the differences in their approaches and outcomes in this case does underline the need for a proper set of regulations governing the manner of any inquiry. N. Raghuram of the SSV talks about a national body analogous to the Office of Research Integrity in the US which lays down requirements for institutional policies and procedures and is empowered to penalize institutions that refuse to act with cuts in funding. It might also not be a bad idea to initially allow the various national institutes to set up their own procedures to deal with such issues and for a national body to be empowered only to review their actions. Comparitive assessments of their experiences might allow a better understanding of the problems and potential solutions inherent in each of them thereby allowing a more uniform and comprehensive system to evolve with time.

Update: Gopal Kundu responded today in The Hindu to the article by Divya Gandhi. He reiterates the evidence and reasons behind his exoneration by the Padmanabhan committee and also indicates that he has republished the same data in a different journal recently (Thanks to Mr. Zubaid who pointed out a significant error in this update which previously ascribed the republication to a different group; it now stands corrected).

2 comments:

Zubaid said...

Just wanted to point out that the data from the JBC paper was republished recently by the same authors - not by a different research group.

Dilip said...

Dear Zubaid,

Thanks for pointing that out. The error now stands corrected.