Functional genomics studies of cancer and cells of origin: from pan-cancer to single-cell

Thesis

The development of stratification systems and therapeutical targets is critical for cancer research. Recent advances in high-throughput sequencing enable the gener- ation of a large amount of cancer genomic and transcriptomic data, which can be used to answer the questions on cancer diagnosis and treatment. However, it is still a great challenge to interpret a mass of data and get reliable conclusions.

In this study, I used the functional genomics methodologies to study cancer and cells of origin. In the pan-cancer study, I developed an algorithm, Masonmd, to predict the targets of nonsense-mediated decay (NMD). This algorithm identified over 140 thousand NMD-elicit mutations from the genomic data of The Cancer Genome Atlas (TCGA) pan-cancer cohort. Further analysis showed that NMD was prevalent across cancer types and, more notably, it could play a permissive role in cancer progression. The results support the notion that the inhibition of NMD is a promising therapeutic strategy.

In the work on single-cell RNA sequencing (scRNA-seq), I first developed a differential- expression-based clustering algorithm, ClinCluster, tailored for clinical scRNA-seq data. The benchmarking analysis demonstrated that ClinCluster outperformed other commonly used pipelines in both simulated and real datasets. I next gen- erated and analysed the scRNA-seq data from human fallopian tubes, which certain high-grade serous ovarian cancer (HGSOC) arose from. The analysis constructed a cell atlas of human fallopian tubes and identified four novel secretory subtypes. Based on the discovery of cell phenotypic repertoire, the deconvolution analysis of the TCGA data refined the molecular subtyping of HGSOC. This gave rise to a robust prognostic predictor, which was validated in eight independent expression datasets. Together, the evidence suggests that the phenotypic heterogeneity in cells of origin can be inherited by tumour cells, which affects patients’ overall survival. It shed light on a refined molecular diagnosis for HGSOC.

Authors: Hu, Z

• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: cancer; functional genomics; cells of origin; fallopian tube; pan-cancer analysis; ovarian cancer; single-cell RNA sequencing
• Subjects: Genomics; Cancer; Bioinformatics; Molecular biology