| The plummeting sequencing cost is leading to a substantial increase in the number of personal genomes. Understanding the functional consequences of the millions of variants obtained from sequencing remains a challenge. Functional genomics and disease studies, such as The Encyclopedia of DNA Elements (ENCODE), Online Mendelian Inheritance in Man (OMIM), and The 1000 Genomes Project, provide an unprecedented opportunity to decipher the pathogenicity of genomic variants. My graduate work has focused upon developing methods to estimate the functional impact of variants through data integration, statistical analysis and software development. In this thesis, we present three computational tools and their applications to cancer genomics. These include: (1) NetSNP, a unified network approach to interpret coding variants; (2) ALoFT, a framework to annotate and estimate disease-causing potential of loss-of-function mutations; (3) a systematic investigation of coding and noncoding mutation patterns in healthy and cancer individuals; (4) FunSeq, a method developed based on knowledge learned from (3) to prioritize regulatory mutations in cancer; and (5) method applications to study noncoding regulatory variants in ∼100 gastric cancer genomes. These methods will be useful for researchers to pinpoint a small subset of diseases-causing variants and could potentially benefit therapeutic treatment. |
