A Genome-Wide Gene-Based Multivariate Phenotype Association Analysis in Families

• Activity Number: 537
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Saonli Basu*+
• Companies: University of Minnesota, Biostatistics SPH
• Keywords: genomewide association analysis ; multivariate phenotype genotype association ; O'Brien's test

Abstract

Implementing genome-wide gene-based association tests for multivariate traits present substantial analytical and computational challenges, especially for family-based designs. Here we propose an extension of OBrien's test to conduct a gene-based association analysis for multiple correlated traits. We first combine the test statistics from single marker association analysis for markers within a gene to produce a gene-based test statistic for each trait, which are then combined to produce a test for association between the multivariate phenotype and the gene, while properly adjusting for the correlation among the phenotypes. We have compared through extensive simulation several gene-based association analysis approaches for multivariate traits. Our method maintained valid type-I error even for genes with markers in strong linkage disequilibrium. It also had substantial power to detect pleiotropic effects. We have also studied their performances on Minnesota Center for Twin Family Research dataset. In summary, our proposed approach provides an efficient and powerful way of conducting a genome-wide gene-based multivariate phenotype association analysis in families.