Omnibus Weighting Incorporating Multiple Functional Annotations for Whole Genome Sequencing Rare Variant Association Studies

• Activity Number: 137 - Statistical Methods for Analyzing Genetic Variants and QTLs
• Type: Contributed
• Date/Time: Monday, July 29, 2019 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Xihao Li* and Zilin Li and Hufeng Zhou and Sheila Gaynor and Yaowu Liu and Han Chen and Alanna C. Morrison and Eric Boerwinkle and Xihong Lin
• Companies: Harvard T.H. Chan School of Public Health and Harvard TH Chan School of Public Health and Harvard University and Harvard T.H. Chan School of Public Health and Harvard TH Chan School of Public Health and the University of Texas Health Science Center at Houston and University of Texas School of Public Health and University of Texas School of Public Health and Harvard
• Keywords: Whole genome sequencing; rare variant association testing; functional annotations; annotation principal components

Abstract

Whole-genome sequencing studies have been increasingly conducted to investigate associations between susceptible rare variants (RVs) and diseases/traits. Variant-set tests are commonly used to analyze RVs in WGS. Existing variant-set tests only incorporate minor allele frequency as weights in RV association analysis. External biological information based on various functional annotations can be utilized to characterize variant function, and thus help boost power for variant-set tests. In this paper, we develop the variant-Set Test for Association using Annotation infoRmation (STAAR), a general framework that incorporates multiple complementary functional annotations as an omnibus weighting scheme to boost power for variant-set tests in WGS studies. First, we propose using Annotation Principal Components (aPCs) to capture multiple aspects of biological function. We then derive the STAAR test statistics to incorporate aPCs in RV association testing. Simulation studies show that the proposed STAAR tests control type I error rates and achieve substantially greater power compared to conventional variant-set tests. We apply the STAAR framework to analyze lipid traits in ARIC WGS study.