Quantifying replication dynamics to predict clonal evolution and drug sensitivity in cancer cells using single-cell whole genome sequencing

点击翻译按钮获取中文摘要

Project Summary Recent developments in single-cell whole genome sequencing (scWGS) such as direct library preparation (DLP) have enabled rich interrogation into the genomic diversity of hard-to-treat breast and ovarian cancers. Studies from the Shah Lab have used DLP to infer clonal fitness from time-series sampling and assess the impact of mutational processes on structural genomic plasticity; however, these studies only looked at genomes from non-replicating cells. The proposed project leverages S-phase cells captured by DLP to prospectively predict clonal fitness and chemosensitivity using data from a single time point (Aim 1) and, separately, to measure the contribution of aberrant replication towards the accumulation of copy number aberrations (CNAs) across cancers of various genetic context (Aim 2). For my first aim, I propose a computational framework for assigning S-phase cells to phylogenetically inferred clones and subsequently calculating the S-phase fraction (SPF) within each clone. For my second aim, I propose a Bayesian model to estimate single-cell replication timing in a manner that is robust to somatic CNAs. I hypothesize that SPF will be a proxy for proliferation rate and thus high SPF clones will have higher treatment-naïve fitness and sensitivity to platinum-based chemotherapy than their low SPF counterparts. I further hypothesize that heterogeneous replication timing will correlate with the extent of replication stress in a particular DLP sample. Investigating both hypotheses will be important for extending the clinical utility of DLP since identifying which clones will respond to chemotherapy will help guide clinical management and identifying the presence of replication stress will help stratify patients for next-generation targeted therapies, such as ATR inhibitors, that selectively kill cancer cells undergoing replication stress. In summary, the proposed computational methods will extend the utility of scWGS technologies which capture replicating cells by connecting replication dynamics to clonal fitness, replication stress, CNA patterns, and drug sensitivities in genomically unstable cancers.