What if scientists could wave a magic wand that would reveal everyone’s DNA, then be able to say, “Aha!—everyone with diabetes has this kind of pattern, and everyone with heart disease has that kind of pattern?” That is the goal of what are known as genome-wide association studies, or GWAS. NIH scientists are currently trying to parse the human genome for disease signatures, and also taking pains to assure that collection of such data maintains donor privacy.

Genome-wide association studies scan a person’s entire genome and explore the connection between genotype and phenotype information to find changes associated with a particular disease or condition. They hold tremendous promise for improving public health. GWAS have the potential to uncover new and more effective methods of prevention, diagnosis and treatment.

What still needs to be determined are the policies for using the data obtained in GWAS supported or conducted by NIH. How should data be shared for further study among researchers? What are the best ways to protect the privacy of people who participate in the studies?

To address these and other questions, an NIH-wide working group was appointed by NIH director Dr. Elias Zerhouni to look at the issues involved, put together a draft of policy and gather feedback from the public before going forward. A town hall meeting was held recently to hear from the public on the policy.

The benefits of GWAS are “new disease understanding and more treatment options,” said Dr. Elizabeth Nabel, chair of the GWAS ad hoc workgroup and director of NHLBI. She defined GWAS as a study of genetic variation across the human genome, “which is designed to identify genetic associations with observable traits such as blood pressure or weight or the presence or absence of a disease or condition.” The data should be shared, she said, because it could improve the health of the public.

Nabel noted that several institutes are already proactively sharing GWAS data, including NEI, NIDDK, NHGRI, NHLBI and NINDS, and said there should be a mechanism to “maximize data-sharing among members of the scientific community,” which, in turn, could lead to a better understanding of the health needs of the public and more effective diagnostic tools and treatments. She outlined several diseases in which GWAS data has already proven helpful, including hypertension, type-2 diabetes and age-related macular degeneration.

The three core elements of the policy for sharing GWAS data, Nabel said, are data management, scientific publication and intellectual property. A key part of managing data is to protect research participants—volunteers who have given informed consent. Therefore, guidelines have been suggested about how best to remove all personal identifiers before the data is submitted to the GWAS depository. The draft policy also proposes rules for investigators’ access to the data, the publication of research findings and intellectual property claims relating to the information.

During the 90-day public-comment period for the policy, 196 responses were received, Nabel said. Key issues raised by the public included ensuring the privacy of research participants, how the repository would be managed and overseen, ownership of data, timing of data release for patenting and data standards, access and security.

Just days before the town hall meeting, NIH announced the launch of a genotype and phenotype (dbGaP) database, which will archive and distribute data from two sources—the Framingham Heart Study and the Genetic Association Information Network—and potentially could be used for other studies producing GWAS data. At the meeting, Dr. Jim Ostell, chief of the Information Engineering Branch, NLM, gave an overview of the new depository to illustrate how GWAS data could be shared.

Following these presentations, two panels of speakers addressed public questions. At the panel on data management, Dr. Francis Collins, director of NHGRI, emphasized that the strong intention of the policy is to “key-code” all of the genetic information given by research participants. The goal is to make sure “genotypes and phenotypes are not associated with any personal identifier such as name, address or Social Security number,” he said, therefore the draft policy outlines ways to eliminate this information.

People have also asked if the genetic data itself could be an identifier, he said, explaining that right now it couldn’t, because it would have to be compared to other data. “Genotype information without any comparison doesn’t reveal anything about personal identity,” Collins said. “The concern is, might there be, especially in the future as more genetic data is collected on individuals…greater opportunity for mischief, where it might be possible for those link-ups to be made.”

Collins said that though “it does seem that at the present time, the risk of that is extremely low, the issue really is about the future.” The reason for the stipulations about user access to the data in the policy, he explained, is that even if all identifying data for participants is removed, “we want users to take this very seriously, as data that they should treat with care.”

Questions from attendees in the first panel discussion included what to do about samples that were submitted prior to requiring participants to provide informed consent, whether there would be training in the approval process for Institutional Review Boards and whether an association found in a participant’s data would be reported back to the participant. Attendees also raised the issues of the possibility of other government agencies having access to the data, the costs of collecting and transmitting it, the possibility of encouraging “altruistic” participants who aren’t interested in maintaining their privacy and enforcement of the agreements investigators make in order to access the data.

The second panel discussion, moderated by Dr. Laura Lyman Rodriguez, special advisor to the director of NHGRI, focused on scientific publication and intellectual property. She said that during the public-comment period, the main questions raised in these areas concerned the 9-month period investigators must wait before publication, what would be done about conflicting conclusions in the literature, if there would be collaboration with the original investigators on new research published and procedures for acknowledgement, as well as issues concerning the ownership of data and the timeline for patentability.

All of the issues raised during the meeting will be considered as the working group continues, Nabel said, emphasizing that they want an “open, deliberative process,” that addresses all concerns. The group hopes to have the final policy in place by this spring, she said, adding that its members “look forward to an ongoing dialogue.”

The draft policy is available online at http://grants2.nih.gov/grants/gwas/.