well in both brain and blood, a notable portion is specific to genes that are only expressed in brain. Also, many SNPs influence multiple molecular features, with a small number of them having their impacts on gene expression mediated through epigenetics. Further, we apply a computational approach to prioritize the cell types that may be driving the tissue-level effect, a critical piece of information for designing follow-up molecular experiments in which an in vitro or in vivo target cell type needs to be selected. Finally, we illustrate the efficacy of an “xQTL-weighted GWAS” approach for applying our xQTLs. We show that this approach increases the statistical power of GWAS, resulting in the detection of a number of new susceptibility variants for several diseases. All data used in this study are available from www.radc.rush.edu, and the xQTL results and analysis scripts can be accessed through our online portal, xQTL Serve, at http://mostafavilab.stat.ubc.ca/xQTLServe.
