ENCODE guided WGS analysis can identify trait associated regulatory regions driven by rare variants

Abstract Large-scale whole-genome sequencing (WGS) data provides unprecedented opportunity to explore the role of rare variants in non-coding regions on complex traits. However, unlike gene-based exome analysis, non-coding regions lack a well-defined unit for rare-variant testing. Here, we utilized 1,036,913 candidate Cis-Regulatory Elements (cCREs) from ENCODE as analysis units and conducted rare-variant association analysis for 100 traits with SAIGE-GENE+. In a discovery set of 150K White-British samples we identified 1,987 significant associations and replicated 88% of them in a validation set of 250K White-British samples. Associations were enriched in promoter-like signals (PLS) and proximal enhancer- like signals (pELS). Conditional analyses of 230 cCREs across five traits on GWAS variants identified 68 independent cCRE associations. A leukemia case study highlighted key loci, including a PLS for SRSF2 and a pELS for BCL6, demonstrating the effectiveness of cCRE- based analysis.