To identify the causal mechanisms leading to disease emergence, endemic maintenance and epidemic expansion,epidemiologists would benefit tremendously from a full characterization of pathogen spread in space and time. The extensivesequencing efforts for the ongoing swine-origin H1N1 influenza A epidemic testify that analyses of pathogen genetic data arebecoming increasingly important source of information in epidemiology. The new era of metagenomics promises to harvestsuch data in unprecedented quantities with remarkable rapidity. With the objective to assist in designing effective interventionand prevention strategies, we propose to develop a comprehensive statistical framework for uncovering the spatial andtemporal dynamics of pathogen genomes. To this aim, we will design and extend a series of computationally tractable modelsthat use the rapidly proliferating viral genome data to their full potential, connect molecular evolution to underlying spatialprocesses, and pave the path to rigorous and powerful phylogeographic hypothesis testing approaches. Armed with thesetools, we will investigate how genetic variation within important viral pathogens, such as influenza, rabies and HIV, arises andpersists across a diverse range of hosts and environments.