Methods for mapping disease susceptibility genes using allele -sharing statistics in linkage analysis and haplotype-based association studies.

Mapping genes for complex human traits is facilitated by two commonly used analytical methods: linkage analysis and association analysis. Linkage analysis is employed to identify broad chromosomal regions containing putative disease susceptibility genes. Association analysis, which tests for smaller regions of linkage-disequilibrium between the disease predisposing locus and nearby marker loci, is used to narrow the candidate chromosomal region containing the putative disease locus. This dissertation consists of three papers that describe alternative approaches for nonparametric linkage analysis and haplotype-based association analysis. In the first paper, I describe the construction of 12 allele-sharing statistics, including the pairs and all statistics presented by Whittemore and Halpern (1994), for use in non-parametric linkage analysis. I compare the relative power to detect linkage between these twelve statistics under specific mode-of-inheritance disease models and show that Statistic A (Lange and Sobel, 1996) clearly outperforms the pairs and all statistics under the proposed recessive disease models and that my two new statistics (S-#alleles dom and Spairs dom ) show promise for mapping dominant genes in pedigrees containing at least two generations of affected individuals. In the second and third papers, I describe a class of nonparametric haplotype-based association. This group of tests, collectively called the haplotype runs test (HRT), is based on conserved runs of shared alleles between pairs of haplotypes. The HRT is designed to be less sensitive than standard likelihood-based tests to the effects of recombination, mutation, and genotype error. The HRT allows inclusion of many markers simultaneously without contributing to concerns of sparse haplotype counts or excess degrees of freedom. I show that the HRT outperforms the popular single locus transmission disequilibrium test (Spielman et al., 1993) and the haplotype-based transmission disequilibrium test (Clayton, 1999) when the number of founder mutations is modestly large. I also illustrate the HRT using previously published data from association studies on the mendialian disorders ataxia-telangiectasia and the Nijmegen breakage syndrome. I conclude with some proposed extensions to the HRT and other suggestions for future research.