Abstract
Contemporary patterns of genetic variation reflect the cumulative history of a population. Population splitting, migration, and changes in population size leave genomic signals that enable their characterization. Existing methods aimed at reconstructing these features of demographic history are often restricted in their temporal resolution, leaving gaps about how basic evolutionary parameters change over time. To illustrate the prospects for extracting insights about dynamic population histories, we turn to a system that has undergone dramatic changes on both geological and contemporary timescales-an urbanized, near-shore archipelago. Using whole genome sequences, we employed both common and novel summaries of variation to infer the demographic history of three populations of endemic white-footed mice (Peromyscus leucopus) in Massachusetts' Boston Harbor. We find informative contrasts among the inferences drawn from these distinct patterns of diversity. While demographic models that fit the joint site frequency spectrum (jSFS) coincide with the known geological history of the Boston Harbor, patterns of linkage disequilibrium reveal collapses in population size on contemporary timescales that are not recovered by our jSFS-derived models. Historical migration between populations is also absent from best-fitting models for the jSFS, but rare variants show unusual clustering along the genome within individual mice, a novel pattern that is reproduced by simulations of recent migration. Together, our findings indicate that these urban archipelago populations have been shaped by both ancient geological processes and recent human influence. More broadly, our study demonstrates that the temporal resolution of demographic history can be extended by examining multiple facets of genomic variation.