Characterization of cancer risk and response to immunotherapy through pathway-based genomic analyses.

摘要

For most cancers, the genetic mechanisms that predispose normal cells to neoplastic transformation and tumor cells to therapeutic eradication are poorly understood. High throughput genotyping technologies have allowed the rapid discovery of associations with these traits in both germline and cancer genomes. Rare or strongly deleterious single-gene mutations, the "low-hanging fruits", tend to be well characterized. However, other genomic variations in the majority of cancer patients still lack adequate interpretation. By considering the effects of many biologically related variations together, i.e. through pathway-based approaches, this dissertation (1) presents a new method of inferring susceptibility mechanisms for cancers of the lung, breast, and prostate that better captures the complexity of tissue-specific protein networks; (2) aggregates risk allele influences of germline markers into pathways and pathway-smoking interactions to more accurately assess individual-level lung cancer susceptibility; and (3) predicts immune response to a novel dendritic cell vaccine among patients with metastatic colorectal cancer using a model of somatic mutations with differential weighting based on biologic pathway involvement.