Using Tumour DNA Methylation for Breast Cancer Subtyping and Prediction of Survival

This thesis explored the role of genome-wide DNA methylation for breast cancer subtyping and predicting survival outcomes. It reviewed the heterogenous literature on this topic. It then assessed the agreement between clustering algorithms for DNA methylation-based subtyping, revealing poor to moderate agreement. Replication analyses undertaken in the Melbourne Collaborative Cohort Study (MCCS) showed partial support for previously reported methylation markers of survival. It created a pooled dataset including close to 2,000 tumours, used to identify individual markers and derive a 228-CpG signature predictive of survival. These findings highlight the promise of DNA methylation to improve breast cancer classification and prognostication.