My work has focused on studying heterogeneity and clonal dynamics in tumors, in particular during the benign to malignant and malignant to metastatic transitions. By combining DMBA/TPA skin carcinogenesis with the Confetti mouse, I have developed a model in which individual tumor clones can be fluorescently labeled and traced, and isolated for analysis. Using this model, I have been able to demonstrate that, while tumor initiation and progression to malignancy are clonal events driven by one cellular population, multiple tumor clones participate in metastasis. This mirrors recent findings of polyclonal metastases in patients, where multiple genetically distinct metastasizing clones have been detected through sequencing. The observation that metastases are polyclonal suggests that metastasis is a fundamentally different process than earlier stages of progression, and opens up the possibility of a role for cellular cooperation. Confetti-DMBA/TPA model, which enables isolation of distinct clones, will be particularly useful for dissecting the process by which these polyclonal metastases arise, including the relative contributions of participant clones and assessment of cooperation between them. In prior work, I was able to show that tumors have intrinsic properties that determine their ability to metastasize, separate from host “readiness” to permit metastasis. Through analysis of phylogenetic relationships between tumors, I was also able to show that metastases diverge in a synchronous parallel fashion from their primary tumor, answering the much-debated and clinically important question of whether or not metastases always progress via a lymph node, which they do not. I am also interested in understanding the mechanisms of interaction between heterogeneous tumors and the immune system. T cells recognize tumor neoantigens, which are the product of genetic mutations, and therefore display the same heterogeneity in the tumor that mutations do, however, this critical aspect of tumor-immune interactions has been largely overlooked.