|
Abstract:
|
Haplotype phase structure may help discover genetic variants associated with disease. However, haplotype-based methods become complex with many loci. Haplotype sharing statistics have been proposed as a solution. These methods consider haplotype similarity instead of modeling each haplotype because, if a recent disease-causing mutation is introduced once into a population, then haplotypes with the mutation should be similar to each other. Thus, one compares the extent of similarity between case and control haplotypes (for case-control data) or transmitted and untransmitted haplotypes (for case-parent data). Existing haplotype sharing statistics use permutation, making them unsuitable for whole genome studies or simulation experiments. We give a class of haplotype sharing statistics and derive its asymptotic properties. A Crohn's disease data set illustrates our approach.
|
