Detecting Joint Effects of Multiple Genetic Variants via Composite Haplotype Approach

• Activity Number: 337
• Type: Topic Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Chia-Ling Kuo*+ and Dmitri Zaykin
• Companies: National Institute of Environmental Health Sciences and National Institute of Environmental Health Sciences
• Keywords: genetic association analysis ; robust statistic ; generalized linear model ; score test ; trend test ; Sime's test

Abstract

Genome-wide association studies have successfully identified thousands of susceptibility loci from marginal association analysis but these variants can only explain a small amount of trait variation. To study undiscovered genetic and environmental factors responsible for the remaining variation, multi-locus association methods with flexibility to incorporate environmental covariates are needed. We define the composite haplotype as a set of alleles across loci, each contributing one allele. The methods we propose for analyzing composite haplotypes use the techniques of frequency filtering and principal component analysis to reduce dimensionality and utilize the regression F-statistic to test for the association with a binary or continuous trait. We show by simulation that compared to a selection of genotype-based and haplotype-based methods, the sum of individual regression F-statistics maintains robust power for a variety of trait models. Since the method is built in the framework of generalized linear models, it is readily applied for various trait types and able to incorporate environmental covariates.