Yanxiao Zhang

Yanxiao Zhang, Ph.D.
16

Ph.D. Program
Post Doctoral Fellow
University of California-San Diego

Chair

Dissertation Title

High-Throughput Bioinformatics Approaches to Understand Gene Expression Regulation in Head and Neck Tumors.

Research Interest

Cancer is defined as uncontrolled growth of abnormal cells bearing various molecular aberrations. With the aid of massively parallel DNA sequencing technology, we can now comprehensively characterize the genomic, epigenomic and transcriptomic landscapes of cancers. Subtypes of cancer are continually being uncovered, often by clustering expression profiles or determining driver mutations, the identification of which can be very important for prognosis and personalized treatment plans. The goal of this dissertation is to develop and apply bioinformatics algorithms to study subtypes of head and neck tumors. Using computational approaches, we both uncover new subtypes, and investigate the oncogenic mechanisms of driving molecular events in the tumor subtypes. My dissertation consists of three main chapters. In the first chapter, we present a software program (PePr) for conducting the differential binding analysis of replicated ChIP-seq data. PePr estimates the biological variation among samples and reports consistent changes across sample groups. We use PePr to characterize the difference in histone modifications between two human papillomavirus (HPV)-associated cancer cell lines and two non-HPV cell lines. In the second chapter, we identify two robust HPV(+) head and neck squamous cell carcinoma subtypes based on gene expression clustering. One subtype (HPV-KRT) shows more frequent genic viral integration and splicing of E6, and reduced viral oncogenic E6 activity. The HPV-KRT subtype also has more frequent copy number gains of chr3q, fewer losses of chr16q, and more PIK3CA mutations. These genomic changes could potentially lead to the differences in gene expression between the subtypes, including elevated immune response and mesenchymal differentiation in HPV-IMU subtype, and up-regulated keratinocyte differentiation and oxidation-reduction process in HPV-KRT subtype. In the last chapter, we characterize the binding profile of a fusion oncogene, PPFP (fusion of PAX8/PPARG; observed in 30% of follicular thyroid cancer) using ChIP-seq data from a rat PPFP-transfected PCCL3 cell line. Our RNA-seq and ChIP-seq results suggest that PPFP regulates many pathways related to cancer, and a PPARG agonist, pioglitazone, may reverse the oncogenic effect of PPFP by altering oxidative stress. Altogether, we demonstrate how the integrative analysis of high-throughput data can guide subtype discovery and mechanistic research in cancer.

Current Placement

University of California-San Diego