Landscape genomics of white-footed mice along an urban-to-rural gradient

Background/Question/Methods

Over 50% of the human population now lives in cities, and urbanization is one of the most important drivers of land transformation around the world.  Increasingly, humans are also a selective force driving rapid evolutionary change in other species.  In this study, we use a landscape genomics approach to examine how urbanization structures both neutral and adaptive genetic variation in white-footed mice.  Using RAD-Seq, we generated high-density, genome-wide SNP (single nucleotide polymorphism) genotypes from over 200 mice sampled from 25 populations along an urban-to-rural gradient spanning New York City to rural Connecticut.  We then used traditional outlier statistics such as F_ST to identify candidate genomic regions experiencing selection, and spatial approaches to identify a subset of SNPs that show a strong threshold in frequency along the urban-to-rural gradient.  We are also using linear modeling approaches to examine environmental correlations between SNP frequencies and variables related to urbanization.

Results/Conclusions

Preliminary results indicate that spatial models based on relatively few high-contrast landscape variables (e.g. vegetation vs. impervious urban surfaces) can explain connectivity vs. abrupt spatial SNP frequency thresholds between urban-suburban-rural populations.  Genomic regions containing coding sequences involved in metabolism, immunity, and reproduction exhibit statistical signatures of selection in isolated urban populations of white-footed mice.  Ongoing work examines variation in gene expression and environment-genotype correlations in urban vs. rural populations of white-footed mice.