A framework for detecting noncoding rare variant associations of large-scale whole-genome sequencing studies

Large-scale whole-genome sequencing studies have enabled analysis of noncoding rare variants’ (RVs) associations with complex human traits. Variant set analysis is a powerful approach to study RV association, and a key component of it is constructing RV sets for analysis. However, existing methods have limited ability to define analysis units in the noncoding genome. Furthermore, there is a lack of robust pipelines for comprehensive and scalable noncoding RV association analysis. Here we propose a computationally-efficient noncoding RV association-detection framework that uses STAAR (variant-set test for association using annotation information) to group noncoding variants in gene-centric analysis based on functional categories. We also propose SCANG (scan the genome)-STAAR, which uses dynamic window sizes and incorporates multiple functional annotations, in a non-gene-centric analysis. We furthermore develop STAARpipeline to perform flexible noncoding RV association analysis, including gene-centric analysis as well as fixed-window-based and dynamic-window-based non-gene-centric analysis. We apply STAARpipeline to identify noncoding RV sets associated with four quantitative lipid traits in 21,015 discovery samples from the Trans-Omics for Precision Medicine (TOPMed) program and replicate several noncoding RV associations in an additional 9,123 TOPMed samples.