| Haplotype, a sequence of alleles on the same chromosome that were inherited as a unit, plays a very important role in genetic studies, such as in disease gene mapping and population genetics. In disease gene mapping, haplotype-based methods have played a major role in the study of Mendelian disorders caused by single genes, and may play a key role in the study of common complex traits as well, suggested by recent studies. However, there are some issues in haplotype association analysis need to be addressed. Among them, one is the effect of missing genotypes, another is the appearance of population structure, and the third is the usefulness of haplotype block and tagging SNPs. This dissertation aims to address these three issues.; In the first part, we evaluate, by simulation, the effect of nonrandom missing genotypes in haplotype frequency estimation and haplotype association analysis. We propose a model to characterize missing data patterns across a set of two or more markers simultaneously. We illustrate that our proposed model can reduce the bias caused by incorrectly assuming missing at random, and have reliable estimates. In the second part, two methods are proposed for population structure inference. One is based on maximum likelihood theory using mixture modeling, where model selection can be naturally incorporated; the other is nonparametric in the sense that no population genetics assumptions are needed. Both methods run quickly and have comparable results with the currently commonly used method. In the third part, several similarity measures are proposed to facilitate the comparisons of haplotype structures, namely block boundaries and tagging SNPs, across populations and from different block partitioning and tagging SNP identification methods. We applied these measures to a real data set on chromosome 10 in 16 worldwide populations and had some interesting findings. |
