Bayesian statistical models for HIV evolution
Alexander Braunstein · Scholarly Commons (University of Pennsylvania) · 2009
Statistical models provide an important mechanism for describing and understanding sequence evolution, such as the escape response of a viral population under a particular therapy. We present a new hierarchical Bayesian model that incorporates spatially varying mutation and recombination rates into a coalescent framework for sequence evolution. Focusing on evolutionary responses to therapy, we maintain separate parameters for treatment and control groups, which allows us to estimate treatment effects explicitly. Our approach is used to investigate sequence evolution at the nucleotide level of HIV populations exposed to a recently developed antisense gene therapy, as well as a more conventional drug therapy. Detection of biologically relevant signals in both studies and recovery of true mutation and recombination rates in extensive simulation studies demonstrate the effectiveness of our method.