"HIV is one of the most rapidly evolving organisms known. Understanding its evolutionary dynamics is essential for successful drug treatment or vaccine design. At the same time, this rapid evolution makes HIV an ideal model system to study fundamental problems in evolutionary dynamics: In HIV, one can directly observe evolution over genetic distances that correspond to millions of years of evolution in other organisms.This proposal combines time series of ultra-deep sequence data of HIV from individual patients, functional information on drug resistance, and methods from statistical physics to study evolution. The sequence data will observe the dynamics of the genotype distribution in the population, while the exceptionally well characterized biology of HIV will allow the assignment of functional significance to the observed genotypic changes. These two levels of description will be integrated by theoretical models that describe how selection on phenotypes feeds back on the genotype distribution.Specifically, we will determine the fundamental parameters of HIV evolution such as selection strength, recombination rates, and the patterns of genetic interactions from the time resolved data obtained by deep sequencing. We will use the data base of viral sequences that evolved in response to drug treatment to infer the fitness landscapes of drug resistance. These two projects will be integrated in a quantitative model of drug resistance evolution. In a third project, we will quantify how genetic interactions affect the formation of circulant recombinant forms of HIV.Using HIV as a model system, we will develop and test theories of multi-locus evolution, characterize fitness landscapes and genetic interactions, and quantify the impact of recombination on HIV evolution."