Tumours are heterogeneous both between patients and within a patient’s individual lesions. We have mined both genetic and non-genetic determinants of clonal heterogeneity from the standpoint of tumour progression, metastasis, and therapy response. By cloning individual components of a basal-like breast cancer model, we showed that individual lineages had the ability to form vessel like structures that are perfused and form vascular mimics. This process is driven by a transcription factor that regulates embryonic vasculogenesis, FoxC2. We have also shown metabolic impacts on metastatic potential based on the availability of non-essential amino acids. More recently, we have developed a new technology, WILD-seq, which provides a much more efficient method to mine tumour clonal heterogeneity, marrying molecular barcoding with single cell sequencing. This has also pointed us to metabolic vulnerabilities in the clonal populations that seed therapy resistance. We hope that this suite of tools will be useful in mining the inherent heterogeneity present within tumour models, even cell lines that have been in culture for many years, for new insights into how to predict patient responses and impact the course of their treatment trajectories.