By Cara Garretson, IDG News Service
April 7, 2002 | WASHINGTON -- The National Academy of Sciences (NAS) is attempting to settle one of biotech’s thorniest issues — the sharing of published works and supporting data.
In the scientific community, publishing is essential to both deepening the collective understanding of a subject and advancing an author's career. What isn't so clear is the author's responsibility to make the data and materials related to a published work available to other researchers.
The NAS took baby steps toward reaching consensus at a one-day meeting held recently by a committee of the academy’s National Research Council.
At the meeting, the Committee on Responsibilities of Authorship in the Biological Sciences offered this draft guideline: It is the author’s responsibility to undertake reasonable effort to make data and materials integral to a publication reasonably and promptly available in a manner that furthers science.
The proposed guideline immediately sparked debate from scientists, researchers, journal editors, and federal agency officials in attendance about the meaning of “integral” and “reasonable.” Committee members pledged to gather more input from the scientific community in the hope of clarifying such details before publication.
"This is not the sum total of what the academy will cover. This is a principle," says Howard Hughes Medical Institute president Tom Cech, who chaired the meeting.
The group hopes to complete the guidelines by late July or early August and to publish them as an NAS report, says committee member Mary Waltham, an independent publishing consultant and the former publisher of Nature magazine.
The guidelines should help determine to what extent published works that underlie an article’s findings, such as brain images or information stored in a genomic database, can be freely shared.
The scientific community has been mulling over these issues for decades, even centuries, Cech says. Recent technological advances, such as the large databases used in genomic sequencing or imaging analysis, have put the debate in the limelight.
The importance of setting guidelines for data- and materials-sharing has increased with the emergence of these databases, Cech says. “That provided a new feature to the landscape — if you publish a paper where you have created a new database and are using that as a main part of your paper or as background, what are your responsibilities as an author?" he asks.
The question of degree has always complicated the sharing of data. Should an author make available only enough data to allow others to verify the findings in the publication, or should all data involved in a publication be accessible? If a software algorithm or database was used to generate the findings, should it also be made available to a requester? And if an author merely describes data in a publication but does not make it available, is the publication a promotional piece for the data rather than an actual finding?
Although many in the scientific community say an author should make materials available upon request for the greater good of science, vexing questions arise regarding the logistics and even the ethics of doing so. For example, who should pay the cost of re-creating and shipping the materials to the requester? What if the author works for a private company and the requester is a competitor? What if the requester turns out to be a bioterrorist? And how can authors ensure that if the materials include human subjects, those identities will be protected?
Other draft guidelines specifically addressed the accessibility of large databases of information and computational algorithms, as well as the protection of human subjects. The committee also attempted to define what happens when an author doesn't comply with the guidelines. The group suggested that the aggrieved first complain to the journal in which the publication appeared, then to the organization the author works for, and lastly to the funding source.
The committee stressed that it is not attempting to set journal editorial policies with these guidelines, but to give the scientific community a framework of accepted standards. Many journal editors at the meeting said they have already devised their own policies for dealing with authors who don't comply with material requests. For example, if an author publishing in the Proceedings of the National Academy of Sciences fails to share materials with a requester, that author cannot publish in the journal again, says Editor-in-Chief Nicholas Cozzarelli.
"It's punitive, but it works," he says.
Yet others at the meeting countered that authors shunned by reputable journals usually can find somewhere else to publish.
Data available on the Web presents another information-sharing problem. In February, a clash developed over data one researcher posted on his public Web site — data that two other researchers cited in their own published piece (see related story XX). The data's owner argued that information found on a Web site can't be considered “published.”
"Now folks say `I believe in this Internet world and I want to disclose my data before publishing ... but don't scoop me on my own project,’ " says Eric Lander, founder and director of the Whitehead Institute Center for Genome Research. A keynote speaker at the conference, Lander says that the community needs a system to protect and encourage posting pre-publication data to the Web.
NAS has a significant task on its hands. If the committee can come up with flexible guidelines that are generally accepted by the scientific community, it could stimulate a freer flow of information. But if the committee's recommendations are viewed as too strict or not in the spirit of scientific discovery, community practices will likely remain in the fractured state they are now.
--Cara Garretson is a Washington correspondent with IDG News Service.