Project Summary/Abstract Cancer causes more than 500,000 deaths a year in the United States, with more than 90% of the morbidity and mortality attributed to metastases. Metastasis is the spread of cancerous clones of cells across a body, whose dispersal routes are not medically detected or visible until later stages of cancer development. These migratory routes are critical to our understanding of the processes that influence tumor diversity in patients as well as aggressiveness, resistance, and escape of tumors from therapy. Comparative analysis of the extensive genetic heterogeneity present in tumors can be used to map the dynamic history of cancer cell migration over time and space. We propose to develop new methods for inferring cancer cell migrations accurately. Our new approaches will employ principles of Bayesian molecular phylogenetics and patterns of mutational signatures in cancer cell genomes for the first time to produce accurate maps of cell migration among tumor sites. We will also develop new methods for detecting mutational signatures. These will overcome many limitations of the current methods for datasets containing a small number of mutations, a situation commonly encountered in single- patient clone phylogenies. Our methodological developments will be complemented by the distribution of a library of functions containing our new and advanced approaches for high-throughput and in-depth analysis of metastatic tumor genomic data. Overall, the proposed software and research developments will lead to advances in cancer, bioinformatics, functional genomics, and data science. New software and its source code will be made available free of charge for all uses, including research, education, and training.