In academic science, ideas and data are exchanged openly. Scientists freely debate the merits of different theories and interpretations of data. Reagents are shared within the community to replicate findings and to further scientific knowledge in new avenues of investigation.

Sounds peachy. But is this reality or is it really naïve fantasy about how science should be done?

Increasingly, academics feel pressure to hide, rather than share, knowledge and materials. Companies that sponsor academic research projects frequently place restrictions on what can be said about the resulting data. And with universities being run more like businesses these days, valuable research materials are often patented or licensed, forcing scientists to pay for access to them.

But a new corporate culture in academia is not entirely to blame for driving researchers to keep their data under wraps. Some degree of competition between fellow academic scientists has always existed. Researchers run the risk of being scooped with their own data if they share too openly. And being beaten to publication is a huge loss for a scientist, as publications are the currency of academic science--they are an important factor in competing for government and private grant funds and in obtaining tenure and respect in a discipline.

Although keeping data private may help the individual researcher, data hoarding is detrimental to the progress of science--or so said 73% of geneticists surveyed in an article in the 23/30 January 2002 issue of the Journal of the American Medical Association. Nearly half of the respondents had been turned down at least once when requesting data or materials in the past 3 years.

On 25 February 2002, a group of prominent academic and industrial scientists, funders, and journal editors gathered at the National Academy of Sciences headquarters in Washington, D.C., to discuss the sharing of scientific data and materials. Prompted in part by Science's decision to publish Celera's version of the human genome sequence without deposition into GenBank, the attendees focused on establishing rules for sharing data and materials referenced in published articles.

National Institutes of Health (NIH) Draft Statement on Sharing Research Data

NIH is accepting comments through 1 June 2002 on its Draft Statement on Sharing Research Data. The final rules will go into effect on 1 January 2003 and will likely require scientists to formally describe in their grant applications how they plan to share the data that will be collected with NIH funding. If applicants feel it is not possible to share the data, they must explain why.

But what rules apply prior to publication? The conventional wisdom in many areas of science is that although data and materials may be shared before a paper describes them, data and materials must be shared afterwards. What happens, however, when a funding agency requires a grantee to release data as soon as it is generated--well before publication? With the advent of large-scale sequencing projects, this is increasingly the case. The National Institute of Allergy and Infectious Diseases, for instance, required Mitchell Sogin, a researcher at the Marine Biological Laboratory in Woods Hole, Massachusetts, to release sequence data for Giardia lamblia as it was being generated by his lab. But now Sogin says that another researcher's group scooped him by using the data in a paper.

All of this prompted Next Wave to ponder sharing in the sciences--what can be done if someone doesn't want to share, and what can be done if someone asks for a resource that is vital and integral to the functioning of one's laboratory? What should be done if data and research materials are discussed at a scientific conference, but are not yet published? Next Wave asked a panel of experts to consider the following fictional case study describing researchers embroiled in a "custody battle" of sorts over an unpublished data set and an antibody:

Lucy Jones, a tenured professor, has pioneered the use of a certain class of drugs to treat AIDS. Despite her skeptics, she has become a well-respected PI who heads a lab with 20 graduate students and postdocs. Jones publishes about 30 papers each year and has received numerous grants.

At a conference, Jones visits a poster presented by Peter Smith, a student in Maxwell Montgomery's laboratory. Montgomery has just set up his lab with a small, 1-year New Investigator Award from a private foundation. Peter Smith is his first and only student right now, but Montgomery was hoping to attract new postdocs by announcing his and Smith's new discovery at the poster session. Their discovery is that one of the drugs in the class championed by Jones inactivates a particular protein called Abc1. This conclusion is based on a huge data set (including thousands of candidate proteins and chemical compounds) originally generated by Montgomery and later added to by Smith. Both also labored very hard to find the precise conditions for administering the drug and have produced the first known antibody to Abc1.

It didn't take Jones long to realize how far the data set and antibody could take her own research--Smith and Montgomery had barely been able to scratch the surface. For instance, one of Jones's postdocs recently found that Abc1 is an activator of cancer tumor growth, but this finding has not been published or presented in public. Jones concludes that the Abc1-inactivator discovered by Smith and Montgomery could also be used to treat cancer, not just AIDS! Once back at her university, Jones e-mails Montgomery asking for the data set and the antibody.

On seeing Jones's e-mail, Montgomery fears that Jones will scoop him with his own data and antibody. He also worries for Smith's thesis and career. If someone else publishes this work, Smith will likely be able to get his Ph.D., but he won't have a publication--not a good situation for a young scientist. Montgomery chooses to ignore Jones's request.

Two months later, Jones hasn't heard from Montgomery. Her postdoc is still unable to find the right conditions for administering the drug, and she is unsure who to credit for the idea of using that particular drug on Abc1 as she drafts her paper. She's also been unable to generate a good antibody to Abc1. Montgomery's data set would make her paper strong, logical, and publishable, and the antibody would be useful for important microscopy studies. But Montgomery seems to be ignoring Jones's e-mail and voice mail messages. Aggravated, Jones is considering putting pressure on Montgomery by contacting the head of his funding organization.

What should Jones, Montgomery, and Smith do in this situation? Here's what the experts had to say:

 

David Gillespie, a postdoc at the University of Utah, wonders if Montgomery might have a "moral obligation to provide a potential treatment of cancer or AIDS. ... Being stingy with his data could cost lives."

 

 

Speaking from the perspective of a private funding organization, Marc Hurlbert, associate director of research at Juvenile Diabetes Research Foundation International (JDRF), says that Montgomery is under no obligation to share until a paper on the data and materials has been published, according to JDRF guidelines.

 

 

An assistant professor at Emory University School of Medicine, Anita Corbett thinks that it is Montgomery's right not to share unpublished knowledge, but concedes that "we do not live in an ideal world--once information is disseminated at a conference, that knowledge is then out in the community."

 

 

Eli Pearce, president of the American Chemical Society, thinks that the team approach is the way to go. He says that Montgomery and Jones should form a joint research program and apply for joint funding.

 

 

Ari Patrinos, associate director of science for biological and environmental research at the U.S. Department of Energy, and Daniel Drell, also with the DOE's office of biological and environmental research, suggest that Jones employ some creative solutions, including "assistance with additional funding sources, perhaps an additional student or postdoc to work in Montgomery's lab (but funded by Jones), an active scientific collaboration, or access to other technologies or resources that Montgomery, on his own, could not command."

 

 

Eric Campbell and David Blumenthal, two authors of the JAMA article on data sharing mentioned above, discuss their landmark study and analyze the implications of withholding data, including the consequences for the next generation of scientists.