Temporal Trends in Results Availability from Genome-Wide Association Studies

The application of genome-wide association study (GWAS) approaches for the study of genetic determinants of common diseases has propelled human genetics forward, resulting in a surfeit of genomic data. With the accompanying level of widespread collaboration and sharing of data, access to this body of valuable genomic data and the application of novel analytic approaches beyond the level of first GWAS scans is yielding additional insights, both in terms of new genetic discoveries and important general biological findings [1]. However, recent work shows that standard statistical approaches can be applied to aggregate genome-wide association results that place individual research participants at increased risks for misuse related to privacy and confidentiality. We define “misuse” as analysis efforts aimed at exposing individual research participants' information, including revealing disease status, predicted future likelihood or past presence of other traits, or attempts to link another DNA result with a participant, for example, to determine presence or absence in a research cohort, ancestry, and relatedness (e.g., paternity/non-paternity). Thus, there is the small but theoretically possible risk of later legal or discriminatory actions that were originally unforeseen by investigators and would likely be unwanted and unexpected by the research participants [2]–[7]. At this time the risks to research participant identification generally exist when there is access to (at least) a moderate number of genetic variant results that include both statistics (regression coefficients or two-sided p-values) and cohort-specific population allele frequencies [2]–[5], [7]. To date, scientific discussion about these potential risks has focused largely on theoretical scenarios and the related ethical and policy responses [2]–[8]. Initial publications [2], [3] resulted in significant policy shifts and reduction in the open access to GWAS results by the creation of controlled access repositories for results (e.g., for the Wellcome Trust Case Control Consortium [WTCCC] results and Framingham Heart Study [FHS] SHARe 100K GWAS results), but the literature contains no systematic assessment of the extent of current GWAS results availability, temporal trends in availability, or the number of studies that remain at a potentially unacceptable level of risk.