Disease-specific prioritization of non-coding GWAS variants based on chromatin accessibility

Non-protein-coding genetic variants are a major driver of the genetic risk for human disease; however, identifying which non-coding variants contribute to which diseases, and their mechanisms, remains challenging. In-silico variant prioritization methods quantify a variants severity in the context of having a phenotypic effect; but for most methods the specific phenotype and disease context of the prediction are poorly defined. For example, many commonly used methods provide a single organism-wide score for each variant, while other methods summarize a variants impact specifically in certain tissues and/or cell-types. Here we propose a complementary disease-specific variant prioritization scheme, which is motivated by the observation that the variants contributing to different diseases often operate through different biological mechanisms. We combine tissue/cell-type specific scores into disease-specific scores with a logistic regression approach and apply it to 25,000 non-coding variants spanning 111 diseases. We show that disease-specific aggregation of tissue/cell-type specific scores (GenoSkyline, Fit-Cons2, DNA accessibility) signifiantly improves the association of common non-coding genetic variants with disease (average precision: 0.151, baseline=0.09), compared with organism-wide scores (GenoCanyon, LINSIGHT, GWAVA, eigen, CADD; average precision: 0.129, base-line=0.09). Calculating disease similari-ties based on data-driven aggregation weights highlights meaningful disease groups (e.g., immune system related diseases and mental/behavioral disorders), and it provides information about tissues and cell-types that drive these similarities (e.g., lymphoblastoid T-cells for immune-system diseases). We also show that so-learned similarities are complementary to genetic similarities as quantified by genetic correlation. Overall, our aggregation approach demonstrates the strengths of disease-specific variant prioritization, leads to improvement in non-coding variant prioritization, and it enables interpretable models that link variants to disease via specific tissues and/or cell-types.