Powerful Haplotype Testing by Means of Hierarchical Linkage Disequilibrium in Genetic Association Studies

• Activity Number: 618
• Type: Contributed
• Date/Time: Thursday, August 7, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Stefan Boehringer*+ and Brunilda Balliu
• Companies: Leiden University Medical Center and Leiden University Medical Center
• Keywords: genetic association ; parametrization ; high-dimensional data ; cumulants ; haplotype

Abstract

Haplotypes represent simultaneous information on several genetic markers. The number of parameters characterizing the haplotype distribution increases exponentially in the number of loci which results in many degrees of freedom in tests based on comparison of haplotype distributions. We propose the use of joint cumulants to develop a new parametrization of the haplotype distribution. Every parameter corresponds to a cumulant which represents a higher order linkage disequilibrium (LD) parameter for each possible subset of a set of loci. Our definition coincides with the well-known definition of LD for pairs of loci. This parametrization induces a hierarchy in the parameters that can be exploited to formulate specific null-hypothesis with polynomial control of the number of parameters by restricting distribution comparisons to subsets of the parameters. We perform simulations and characterize scenarios where power is improved and show that these scenarios are relevant in practice. We analyze data from the WTCCC study, a Genome Wide Association Study on Rheumatoid Arthritis.