Cancer is an evolutionary process. Mutations in the genomes of cancer cells generate diversity and differential fitness among cells. One implication of this is that most tumours consist of multiple clonal sub-populations of cells. High throughput sequencing (HTS) has emerged as a powerful tool to study cancer genomes. I will present three methods which leverage data from different HTS assays to provide complementary information about the clonal population structure of a tumour. First, I will present a phylogenetic method which uses bulk whole genome sequencing (WGS) data from multiple samples to infer the genotypes of the dominant clones in each sample. Our approach accounts for stochastic under-sampling due to the relatively low coverage (30x-50x) of WGS and allows for point mutation loss due to coincident copy number change. Second, I will discuss the PyClone model which uses targeted deep sequencing data to allows us to infer what proportion of cells in a sample harbour a mutation and which mutations originate at the same point in the evolutionary history of tumour. Third, I will present the single cell genotyper (SCG) model which can be used to analyse targeted single cell sequencing data of known point mutations. The model accounts for several sources of noise, including allele drop-out, and infers both the clonal genotype of the cells and the proportion of each clone in multiple samples. I will finish with some recent results from a study tracking the migration of clones in the peritoneal cavity of high grade serous ovarian cancer patients. We use the three aforementioned methods to determine if the pattern of spread in a patient is indicative of mono-clonal or poly-clonal spread.
Keywords: cancer, genomics, Bayesian statistics, high grade serous ovarian cancer, phylogenetics, single cell sequencing, high throughput sequencing.