| Genetic tools have been instrumental in uncovering the etiology of countless heritable diseases. With the completion of the Human Genome Project and the beginning of its sequel, the Human HapMap Project, the field of genetics is on the verge of being able to solve heritable diseases of an even greater order of complexity. In the opinion of many in the field of gene mapping, the tool which will usher in this new era is the association study.; Currently, no generally accepted method for analyzing genomewide disease-marker association data exists. In the field of gene mapping, it has been the feeling among many researchers that haplotypes, i.e. variants of close physical proximity inherited together, can be exploited to allow greater power in detecting disease mutations. How haplotypes can be obtained from the results of commonly-employed sequencing methods itself presents a challenging problem. Tasks further down the analysis pipeline of detecting disease association---what statistical methods to use, how to correct for multiple tests, computer run-time considerations, and in what circumstances haplotypes are useful (or detrimental)---all require further investigation, too.; This thesis lays out a series of computer algorithms addressing all the points above. First, a computational method is described to infer haplotypes among samples of diplotypes under a neutral coalescent assumption. The results from application of the program confirm that in samples of unrelated individuals, there is insufficient information to phase accurately across areas of low linkage disequilibrium. To compensate for this deficiency, an extension of the original computational method is described to allow the incorporation of familial genotype information. Finally, a method for carrying out family-based association analysis using haplotypes is presented. The program is efficient and adjusts for multiple tests. Applications to simulated data suggest that association studies typing hundreds of thousands of SNPs and analyzed by the methods described herein will be remarkably powerful. |
