Tuesday, January 10, 2006

Implications of the Hwang case for authorship in academia

Seoul National University released its final report (official English summary here) on the fabrication of results by discredited stem cell researcher Woo Suk Hwang. I'm not a molecular biologist, and am even further from being a specialist in stem cell research, so I'm going to refrain from attempting to summarize exactly what has happened biologically in this case. However, the case is interesting, so here's a short summary of the final report:

1) The results reported in Hwang's 2005 Science paper (Hwang WS, Roh SI, Lee BC, Kang SK, Kwon DK, et al. 2005. Patient-specific embryonic stem cells derived from human SCNT blastocysts. Science 308: 1777-1783) were completely fabricated:
The data in 2005 article including test results from DNA fingerprinting, photographs of teratoma, embryoid bodies, MHC-HLA isotype matches and karyotyping have all been fabricated. ... In conclusion, the research team of Professor Hwang does not possess patient-specific stem cell lines or any scientific bases for claiming having created one.
2) The results in Hwang's 2004 Science paper (Hwang WS, Ryu YJ, Park JH, Park ES, Lee EG, et al. 2004. Evidence of a pluripotent human embryonic stem cell line derived from a cloned blastocyst. Science 303: 1669-1674) were also entirely fabricated:
The claim in 2004 article that the DNA fingerprinting pattern of NT-1 and that of the donor A match perfectly was a clear false report. Given that none of the alleged NT-1 derived cells or tissues match the donor A, the committee concluded that NT-1 ES cell line reported in Science in 2004 is not an ES cell line derived from a cloned blastocyst. In addition, claims that photographs of cells in 2004 Science article are those of MizMedi ES cells have also been confirmed to be true. Therefore, the committee concluded that results described in 2004 Science article including DNA fingerprinting analyses and photographs of cells have also been fabricated.
3) The results in Hwang's 2005 Nature paper reporting on the cloning of a dog (Lee BC, Kim MK, Jang G, Oh HJ, Yuda F, et al. 2005. Dogs cloned from adult somatic cells. Nature 436: 641) were confirmed; the dog exists, and appears to have been cloned.

4) Not only did Hwang fabricate research, but it also appears that he used eggs donated by his junior researchers and technicians, opening questions about whether implicit or explicit coercion was involved:
Regarding the article in 2004, Professor Hwang claimed to have been unaware of the egg donation by the laboratory members. However, the graduate student who donated eggs informed the committee that the act of donation, while voluntary, was approved by Professor Hwang. Egg aspiration was carried out by Dr. Sung Il Roh on March 10 of 2003 at MizMedi Hospital, and notably, Professor Hwang accompanied the student to the hospital himself. In May of 2003, Professor Hwang's research team circulated a form asking consent for voluntary egg donation and collected signature from female technicians.
Science has a very good summary (free full-text) of how this fraud was uncovered by anonymous message board posters and journalists using somewhat unethical techniques.

One of the issues that this case brings up is how authorship on scientific papers is determined. Authorship is typically not determined by who writes the paper (though the person actually writing the text of the paper is typically an author). Instead, the list of authors is determined by who made significant contributions to the research. Often, the only contribution made by authors may have been financial or physical resources (getting grants or providing facilities), or a highly specialized technique that is relevant to only one small portion of the research (and thus the author may have no knowledge of the rest of the work in the paper). In fact, the list of authors on a paper can be determined before the research has even begun; this is often recommended so that it's clear to everyone who's getting the credit for the work (so three people don't argue at the end of the research that they should all be first author), and to encourage people to work together (the idea being that if you know you're going to be an author on a study, you'll work harder).

In theory all authors should review the final manuscript and have a say in what's written, but I have a hard time believing that papers with dozens of authors can have feedback from everyone incorporated into the paper, much less have everyone actually verify that the data were collected properly. Hwang's 2005 Science paper had 25 authors spread across multiple continents. Was it even possible that they could have verified the experimental results, especially given that one of the researchers was lying? And what happens to Hwang's 24 coauthors now that the paper has been shown to be fabricated?

Hwang is not alone in having large numbers of coauthors; the best example I know of is a landmark 2004 ocean fertilization study (Coale et al. 2004) that had 48 authors, which was then cited by another paper (Armbrust et al. 2004) that has 45 authors. Even a quick review of the Science issue Hwang's paper was published in shows that more than half of the research reports had at least five authors, and two (excluding Hwang's) had more than ten authors.

The reason why authorship is handed out so readily is that authorship is academic gold, especially in big-name publications like Science and Nature. Getting a paper in Science or Nature is a Very Big Deal; it gives you the ability to obtain funding (and get positions) you otherwise couldn't get. When there are such huge rewards for obtaining authorship, and no clear guidelines as to what constitutes a contribution sufficiently large enough to obtain authorship, it's no surprise that huge lists of authors are a reality.

Complicating the matter is that a lot of scientific writing is based on trust. To use myself an an example, all of my published papers have been coauthored with at least two other researchers, yet in all of the cases where I was the experimenter carrying out the research, I was the only one of the authors who actually saw all the raw data and analyzed it. The only thing my coauthors saw were the summary graphs I showed them, and me in the lab working hours upon hours at my experimental setups. I easily could have faked the graphs and statistics after collecting the data, and my coauthors would have had no way of knowing, short of re-analyzing tens of thousands of datapoints or redoing months of experiments (especially because, unlike Hwang, all of my experimental organisms were destroyed at the end of the study due to the data collection techniques involved). My coauthors all contributed significantly to the ideas in my papers, and thus deserved authorship, yet they had no realistic way of knowing whether I forged my data or not. So, should my coauthors be held accountable if it were found that I had forged the data?

One possible way to get around some of these issues would be to require authors to list exactly what they contributed to the paper. That way it would theoretically be clear who did what, and then if one portion of the research was found to be improperly conducted, it would be clear who should be held accountable.

No comments: