Data Sharing and dbGaP: A Survey of Practices and Opinions Among Human Geneticists. D. Kaufman, J. Bollinger, R. Dvoskin Genetics & Pub Policy Ctr, Johns Hopkins Univ, Washington, DC.

   Background: Data sharing has become important in human genomic research; the NIH requires that GWAS data generated by federally funded research be deposited into the database of Genotypes and Phenotypes (dbGaP). However a number of challenges have caused frustration among researchers who wish to effectively share resources. Methods: An online survey was conducted to explore current practices and opinions related to data sharing and dbGaP. Participants were ascertained through the ASHG membership (n=159) and NCBIs lists of dbGaP data users (n=169) and contributors(n=29). Results: Three quarters of all respondents (n=357) share genetic and phenotypic data with other researchers. A total of 64% said the use of other researchers samples and data has become more important to their work over the past five years, while 17% said sharing their own data conflicted with their career advancement. Two-thirds agreed that ensuring researchers have fair access to shared samples and data is a problem, while 57% said they thought data or samples are occasionally (44%) or frequently (13%) shared inappropriately in genetic research. Among ASHG members, 38% had contributed to dbGaP and 43% had used dbGaP. Among dbGaP users, 89% felt it had benefited their research; however, 55% said the dbGaP application process was more difficult than expected, and 27% said the data quality was lower than expected. Most (84%) believed that dbGaP contributors should follow a single standard for submitting phenotypic data. Despite these problems 84% supported requiring federally funded researchers to submit data to dbGaP. If it were up to them, though, 23% of researchers and 25% of dbGaP contributors would not submit data to dbGaP. When asked about needs and guidance related to data sharing, 80% said template consent language explaining data sharing through dbGaP would be helpful, and 74% said there is a need for guidance on the preparation of data sets for sharing. Conclusions: A genomic studies begin to require larger sample sizes, access to others data and samples becomes critical. Issues of fair access for all researchers, and sharing while maintaining respect for and privacy of human subjects may escalate in importance. Researchers feel dbGaP can improve, but majorities support the role it plays and the requirements to deposit data, and feel that it has benefited their work. Standards for data submission may be among the most practical needs for researchers today.

You may contact the first author (during and after the meeting) at