Data-driven discovery of cancer therapy targets in the “Dark Matter” of the human genome

Somatic mutations in the non-coding regions of the tumour genome are increasingly recognised as significant contributors to tumorigenesis. Long non-coding RNAs (lncRNAs) are frequently mutated in the non-coding genome and represent promising new targets for cancer therapy. Yet, there is limited evidence that mutations in lncRNAs can act as “drivers” that promote cell fitness during tumorigenesis. This is primarily due to the limited number of driver-lncRNAs identified in statistically underpowered studies, combined with a lack of comprehensive functional investigations, which has restricted our ability to establish a clear link between somatic mutations in lncRNAs and cell fitness. This thesis aims to advance our understanding of the role of somatic mutations in driving lncRNA-mediated tumorigenesis in two parts. First, a multitude of novel cancer-driver lncRNAs were identified by analysing 12,631 tumour whole genomes from the Genomics England consortium. This comprehensive resource of driver lncRNAs enables further in-silico investigation of their distinct biological characteristics. Second, this thesis leverages CRISPR-Cas9 technology to construct a comprehensive library of patient-derived lncRNA mutations, enabling the functional characterisation of thousands of somatic mutations in the context of tumorigenesis. This will enable the generation of a first-of-its-kind fitness-mutation map for lncRNAs. Thus, this thesis presents the most extensive catalogue of cancer-driver lncRNAs to date and lays the groundwork for future large-scale functional screens to elucidate the biological significance of lncRNA mutations. The identification of a broader set of cancer-driver lncRNAs, combined with a deeper understanding of their role in driving tumorigenesis, will facilitate the development of innovative lncRNA-based therapeutic strategies against cancer.