GWAS Data
GWAS Central provides a comprehensive collection of summary-level genetic-association data and advanced visualization tools to allow comparison and discovery of datasets from the perspective of genes, genome regions, phenotypes, or traits.33 The project collates association data and study metadata from many disparate sources, including the National Human Genome Research Institute GWAS Catalog,35 and receives frequent data submissions from researchers who wish to make their research findings publicly available. All gathered and submitted data are extensively curated by a team of post-doctoral genetics researchers who manually evaluate each study for its range of phenotype content and apply appropriately chosen MeSH terms. As of December 2014, the resource contained 69 million p values for over 1,800 studies.
Data and metadata for up to 1,000 associations can be freely downloaded from the BioMart-based system (GWAS Mart), and larger custom data dumps (up to and including the complete database) are available via contacting the GWAS Central development team and agreeing with a data-sharing statement. Thus, to provide data for the present study, we generated a tab-separated file representing 1,574 studies and 34,252 unique SNPs (annotated to 675 unique MeSH terms) and containing the GWAS Central study identifier, PubMed identifier, dbSNP “rs” identifier, p value, and MeSH identifier for all associations with p < 1 × 10−5. We compiled the list of genes considered for our experiments by retrieving the “mapped genes” column from the database SCAN and identifying those genes corresponding to the GWAS Central SNPs. Where no mapped genes were reported, we used the upstream, as well as downstream, genes listed by SCAN.44