Rare Variant Association Analysis of Quantitative Traits in Pedigrees of Arbitrary Size and Structure

• Activity Number: 651
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Yunxuan Jiang* and Michael Epstein and Karen Conneely
• Companies: Emory University Rollins School of Public Health and Emory University and Emory University
• Keywords: Rare Variant ; High Dimension ; Sequencing Study

Abstract

Family-based rare-variants sequencing studies have attractive characteristics: they can be more powerful compared to case-control and population-based studies and can further enable the implementation of rare-variant association tests that, by design, are robust to population stratification. However, current methods for family-based studies mainly focus on nuclear families or parent-offspring trios while ignoring the information provided by other types of relatives. With this in mind, we propose a powerful rare-variant test for analysis of quantitative traits in extended pedigrees that can accommodate any family size, and is robust to population stratification. For each non-founder, our method partitions rare variants in a region of interest into a between-family component and a within-family component. We then test for association between the robust within-family component and the trait of interest using a variance-component score test that accounts for relatedness among family members. Our method maintains power when a region harbors variants acting in different directions on outcome, accommodates covariates, and efficiently calculates p-value using an asymptotic distribution.