Charting the alternative splicing vulnerabilities of cancer

Cancer co-opts alternative splicing to select isoforms that promote disease initiation, progression, and persistence. Drugs targeting either cancer driver splicing factors or specific splicing events can hamper the disease. However, the systematic identification of splicing factors and the splicing events with cancer driver potential –and thus therapeutic potential–remains challenging. In this thesis, I develop computational methods to prioritize splicing factors and events of functional relevance to cancer. I create methods to chart the splicing programs underlying heterogeneous cancer samples, revealing their involvement in multiple cancer hallmarks. Additionally, I establish a statistical framework to identify potential cancer driver exons across the genome enlarging the list of potential cancer driver genes. This approach uncovers exons with therapeutic potential for treating glioblastoma in genes previously unexplored and identifies cancer-specific exon-exon splicing junctions that could drive cancer. Overall, this work provides tools to study the systemic regulation of alternative splicing in cancer and generates functionally relevant, testable hypotheses, to exploit cancer’s splicing vulnerabilities therapeutically.