This dissertation investigates the evolution of the canine transmissible venereal tumour (CTVT), a sexually transmitted cancer which manifests as genital tumours in dogs. This cancer first arose in an individual dog several thousand years ago, and has since survived by transfer of living cancer cells to new hosts during coitus. Today, CTVT affects dogs around the world, and is the oldest and most prolific known cancer clone. As such, it provides a unique opportunity to explore the evolution of cancer over the long term, and to track the unusual biological transition from somatic mammalian cell to asexual parasite. Through computational analysis of the mutations found in the protein-coding DNA of 546 globally distributed CTVT tumours, this work aims to illuminate the history, diversity, mutational exposures and evolution of the CTVT lineage, and to characterise the major evolutionary forces which operate on cancer over long time frames.

The dissertation is structured in seven chapters, each composed of four sections. The first chapter introduces the concepts of tumour evolution and clonally transmissible cancers, with special emphasis on CTVT, and exposes the major research questions and experimental rationale of the study. The second chapter describes the identification, annotation and preliminary analysis of the genetic variation found in CTVT tumours and their host dogs. The third chapter presents phylogenetic and phylogeographic analyses, including the reconstruction and dating of the CTVT tumour phylogeny and the inference of past geographical movements. The fourth chapter explores signatures of mutational processes in the CTVT genome, which include exposure to ultraviolet light and a previously undescribed ancient mutational process, whose activity ceased approximately 1000 years ago. The fifth chapter describes analyses of patterns of gene expression in CTVT cells and tumour-infiltrating host cells, which indicate that gene expression in tumour cells is largely centred on cellular maintenance and replication. The sixth chapter explores signatures of natural selection in CTVT, including early ‘driver’ mutations and signals of selection for more recent mutations; the results reveal that CTVT largely evolves via genetic drift, and has apparently ceased to adapt to its environment. The closing chapter presents the conclusions and future directions of the research.

Overall, this work charts the natural history of a unique biological entity, highlights the potential for long-lived clonal organisms to act as biomarkers for mutagenic exposures, provides insight into the forces which dominate cancer evolution over the long term, and suggests an evolutionary limit to the existence of naturally arising transmissible cancers.