| The Human Genome Project and other related projects have greatly stimulated research efforts to understand the genetic mechanisms and their interactions with environmental factors on traits of great interest both in humans and in other organisms. In order to achieve these goals, the genes that affect these diseases must first be either mapped to a certain region in the genome through linkage studies or identified through association studies. Then a thorough understanding of the joint effects of these genes on phenotypes needs to be fully investigated through statistical and biological analyses. There are many statistical challenges in this endeavor, especially with the availability of DNA sequences as well as millions of variants in different organisms. In this dissertation, we develop statistical methods and compare the power of alternative designs to address scientific problems arising in current genetic research. We first investigate the feasibility and relative efficiencies of different designs to detect gene-gene interactions in humans, both through association studies and through linkage studies. We then evaluate the use of DNA pooling strategy for haplotype inference in population genetics studies. Lastly, we develop novel statistical methods to map genetic loci underlying malaria response in mosquitoes that has great implication in public health. |
