View poster materials
William DeWitt1,2, Kameron Decker-Harris1, Kelley Harris1,2
1) University of Washington, Seattle; 2) Fred Hutchinson Cancer Research Center, Seattle.
Models in evolutionary genetics typically assume that mutation rate is constant over time and between populations and closely related species. However, recent work casts doubt on this assumption in human and ape populations, and reveals that mutation is a complex and dynamic process. Whether arising from variation in replication fidelity, life history, or environmental exposures, mutation rate evolution can be accompanied by changes to the mutation spectrum: the mutation rate in different local nucleotide contexts. We extend theoretical tools based on Kingman’s coalescent to accommodate a richly parameterized mutation process, varying in time and in spectrum. We infer human mutation spectrum histories from patterns of modern genomic diversity, allowing us to reconstruct trajectories of mutation spectrum divergence between populations, track a transient mutation spectrum perturbation through multiple populations, and characterize how mutation spectrum histories are structured by local nucleotide context. Mutation rate and effective population size together determine the strength of genetic drift shaping genomic variation. We introduce fast nonparametric joint inference of mutation spectrum history and demographic history from unphased genomes.