Date of Award

12-17-2020

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Bioinformatics

First Advisor

Mary Yang

Abstract

Cancer is the second leading cause of death in the US. Enormous efforts have been devoted to understanding and treat the disease. Overall cancer mortality rate declines in recent years. Meanwhile, the common biological functions and regulatory mechanisms of the general development of cancer remain to be elucidated. Here, we developed computational approaches to integrate multilayer genomic data for better understanding and characterizing cancer. First, utilizing high-through whole-exome sequencing and RNA sequencing data, we developed a systems biology method that combined somatic mutations and gene expression to uncover biomarkers for breast cancer survival prediction. Our method overlaid somatic mutations on the pathways which effectively improved prediction accuracy and offered new putative therapeutic targets for prolonging patients’ lifespan. Moreover, we established a comprehensive study to investigate recurrent somatic mutations in 33 cancer types for over 10,000 patients utilizing sequencing data from The Cancer Genome Atlas (TCGA). We found that seven top mutated genes, which consisted of well-known as well as novel cancer genes, exhibited mutually exclusive mutation patterns across many cancer types. Additionally, three key signaling pathways (p53/wnt/pik3a-akt signaling pathway were shown to be significantly affected by these mutations ((adjust p < 0.05). LRP1B, one of the seven mutated genes, presently is not involved in any KEGG pathways. Our protein-protein interaction and expression correlation analysis suggested that LRP1B is associated with multiple genes in Wnt signaling pathway. We constructed an interaction network of LRP1B to infer its regulatory roles in various types of cancer. We further extended our investigation to epigenomic analysis and studied DNA methylations in 23 cancer types. Integrating the transcriptome profiles, we identified 217 upregulated and 875 downregulated genes that were associated with aberrant methylations in their promoter regions. These genes were enriched in multiple key biological processes. Furthermore, our comparisons between the recurrently mutated genes and the methylation-enhanced/silenced genes indicated the two independent disease mechanisms. Our project expands the catalog of biomarkers and driver events in cancer and provides novel insights into the molecular mechanisms underlying cancer development. The comprehensive pan-cancer study helps us to perceive the connections among different cancer types and contribute to cancer diagnosis and treatment.

Download

Share

COinS