Development of Novel Methods to Study Chromosomal Instability and Genetic Heterogeneity for Biomarker Identification

Abstract

Genomic instability is a hallmark of cancer, driving tumor evolution and enabling the acquisition of additional cancer hallmarks such as evasion and ability to metastasis. While genomic instability facilitates tumor adaptation and progression, it also makes tumors susceptible to therapies targeting DNA repair pathways (e.g., PARP inhibitors) and immune checkpoints inhibitors. Reliable measurement of genomic instability is essential for predicting patient responses to specific treatments, determining prognosis, and discovering novel therapeutic targets. Despite its clinical significance, there is a scarcity of robust methods to quantify the degree of genomic instability in tumors, leading to a limited number of biomarkers predicting genomic instability and hence patients' responses to treatment in current clinical use. In this dissertation, I present novel computational models to measure genomic instability in tumors. Utilizing these methods, I identify and validate new prognostic biomarkers in Lung adenocarcinoma based on the expression of a specific gene, making them practical biomarkers for predicting genomic instability in clinical applications. Furthermore, I will present a genomic alteration in high grade serous ovarian carcinoma which can be used as a predictive biomarker for patient response to platinum-based treatment. Additionally, I explore the downstream effects of genomic instability, particularly the resultant genomic heterogeneity in prostate cancer. In this dissertation, I aim to enhance the understanding of genomic instability and its implications in cancer biology by providing new computational methods to measure this phenomenon, ultimately contributing to biomarker development and improved clinical outcomes for patients. I first introduce two new measurements of genomic instability for chromosomal changes called Genome Contiguity Index and breakage intensity and clustering. Then using these methods I identify IGF2BP3 as a prognostic biomarker for genome instability in Lung adenocarcinoma. Next, I will show loss of heterozygosity (LOH) of chromosome 17 is associated with a patient's response to platinum-based treatment in high grade serous Ovarian carcinoma. xx Furthermore, I will investigate the underlying biology of chromosome 17 LOH and identify the presence of TP53 functional copy to partially account for the association of chromosome 17 LOH and patient response. In the next chapter, I will investigate the downstream effect of genomic instability and resulting genomic heterogeneity in prostate cancer (PCa). I will quantify the extent of genomic heterogeneity at mutation and CNV level, with a focus on pathogenic alterations in PCa patients. I will also investigate the effect of this heterogeneity on detection of alterations using a single sample, which is common in many of the current studies in the field. Lastly, I will introduce another method for detection of copy number variations (CNV) changes, specifically in Human leukocyte antigens (HLA) genes. Loci of HLA genes has the highest polymorphism in the genome and hence, it requires special methods to identify CNVs in this region. In this chapter, we show that by utilizing dynamic reference genome selection (as opposed to standard reference genome), we can significantly improve our CNV calls for HLA genes.