UCSF

In the first chapter of this dissertation, I summarize previously published findings from genome-wide association studies (GWAS) of 172 specific chronic human diseases. Using this data, I estimate the percentage of additional heritability that could be explained by future GWAS using microarrays with improved coverage of germline single nucleotide polymorphisms versus larger sample sizes (more individuals genotyped). In the second chapter, I describe a method for identifying and visualizing genetic pathways that have a significant number of small somatic mutations in human tumor genomes. This method uses random forests to identify the significant pathways and interactions between them, and hive plots to visualize these results. In the third chapter, I describe the results of my analysis of whole genome sequence data from 24 patients with aggressive prostate cancer. I apply the pathway analysis method described above to the small somatic mutations, and look for patterns of copy number aberrations, structural rearrangements, and differential methylation in these tumor genomes compared to patient-matched normal genomes.