PROJECT SUMMARY The rate at which DNA mutates ultimately determines how many people are born with serious genetic dis- eases, as well as how long a person is likely to live before getting cancer. It is also crucial to understand how mutations generate genetic variation in order to accurately infer evolutionary history from genomic data. Despite this fundamental importance for human health and disease, we know little about how the mutation rate varies from person to person and what genetic factors might cause the mutation rate to vary. My previous research has shown that mutations from different populations are biased to occur in different sequence contexts; for example, Europeans contain more mutations in the motif “TCC” than Africans or East Asians do. This implies that each population is affected by a distinctive combination of sequence-biased mutational processes. Unless these dif- ferences are all induced by environmental mutagens, some of them must be the signatures of “mutator alleles,” genetic variants that subtly affect the likelihood of DNA damage or the efficacy of DNA repair. This proposal describes a multi-pronged strategy for interrogating the causes and consequences of variation in DNA replication fidelity. The first step will be to look beyond short, three-letter motifs to identify longer DNA sequences that differ in mutability between populations. To achieve this, we will adapt statistical techniques that have recently been used to identify the motifs that drive hypermutation in immune cells. Once we identify such motifs, we will scan them for concordance with the rich libraries of motifs that are known to regulate protein binding and gene expression. For the first time, we propose to incorporate ancient DNA into our analyses of human mutation spectrum variation, aiming to improve our understanding of the pace of mutation rate evolution and interrogate the role of global mi- gration events in spreading mutator alleles. As a complement to this work on humans, we will also study mutation sequence context variation in polar bears and brown bears, which have been hybridizing for thousands of years in a unidirectional way with polar bear migrants entering the brown bear population but never the reverse. By analyzing the covariance of mutational sequence context in polar bears and brown bears across a range of allele ages, we will infer how often mutator alleles have crossed from one species to another. In natural populations, is difficult to map the genomic locations of mutator alleles because they are predicted to quickly recombine away from mutations they create; to overcome this difficulty, we are working with collaborators to study mutagenesis in model organisms, where inbreeding can be used to force mutations to stay linked to the genetic backgrounds on which they arise. We will develop methods to map mutator alleles in two different inbred model systems: the BXD recombinant inbred mouse strains and the Drosophila Genome Reference Pane...