Across two continents: the genomic basis of environmental adaptation in house mice (Mus musculus domesticus) from the Americas

Abstract

Parallel clines across environmental gradients can be strong evidence of adaptation. House mice (Mus musculus domesticus) were introduced to the Americas by European colonizers and are now widely distributed from Tierra del Fuego to Alaska. Multiple aspects of climate, such as temperature, vary predictably across latitude in the Americas. Past studies of North American populations across latitudinal gradients provided evidence of environmental adaptation in traits related to body size, metabolism, and behavior and identified candidate genes using selection scans. Here, we investigate genomic signals of environmental adaptation on a second continent, South America, and ask whether there is evidence of parallel adaptation across multiple latitudinal transects in the Americas. We first identified loci across the genome showing signatures of selection related to climatic variation in mice sampled across a latitudinal transect in South America, accounting for neutral population structure. Consistent with previous results, most candidate SNPs were in regulatory regions. Genes containing the most extreme outliers relate to traits such as body weight or size, metabolism, immunity, fat, and development or function of the eye as well as traits associated with the cardiovascular and renal systems. We then combined these results with published results from two transects in North America. While most candidate genes were unique to individual transects, we found significant overlap among candidate genes identified independently in the three transects, providing strong evidence of parallel adaptation and identifying genes that likely underlie recent environmental adaptation in house mice across North and South America.

Keywords: adaptive evolution; climate; environment; parallelism.

Figure 1.

a) Map of mean annual temperature across the Americas (WorldClim database). Populations of wild house mice sampled across a latitudinal transect in Mexico and South America localities are shown with large circles. Populations included in previously published surveys in North America (Phifer-Rixey et al. 2018; Ferris et al. 2021) are shown with small circles. b) Admixture plot including representatives from all three primary subspecies of house mouse as well as mice from sampled populations in the Americas. c) Phylogenetic reconstruction of Mus musculus domesticus populations across the Americas, with M. spretus as the outgroup.

Figure 2.

a) Admixture plot from South American populations evaluating K=2, 5, and 8. b) Genetic PCA of Mus musculus domesticus populations across the Americas. c) Heatmap of pairwise genetic differentiation (F_st) values between the eight populations in South America.

Figure 3.

Signatures of selection among house mice from South America. a) The distribution of outlier SNPs across predicted variant effect categories. Proportions shown are averages across LAT, MAT, and PDM; full data are given in Table S7. b) Venn diagram showing the number of unique and shared candidate genes in the South America transect for latitude, mean annual temperature (MAT), and precipitation of driest month (PDM). c) Manhattan plots showing the results of the population genomic scans for selection for the three environmental variables (red line indicates q-value = 0.05, and blue line indicates a q-value = 0.001). Highlighted are the first 100 hundred SNPs with the lowest p-values for each environmental variable.

Figure 4.

Venn diagram illustrating the shared candidate genes across the three transects [South America (SA), Eastern of North America (ENA), and Western of North America (WNA)], and the significantly enriched gene ontologies for each environmental variable: a) latitude, b) MAT and c) PDM (complete enrichment information is on Supplementary Table S11). CF (Classic fisher test used for the enrichment analysis).

Figure 5.

Body weight and genes associated with body weight in house mice from North and South America. a) Boxplot shows the distribution of adult body weights for each population from Eastern North America (magenta), Western North America (grey), and South America (green). Pregnant females have been excluded. b) Manhattan plot showing the genome-wide association results using exome and body weight data. Significant candidate SNPs (FDR ≤ 0.05) associated with body weight are highlighted in purple with the gene annotated for the SNP (red line indicates p-value = 0.01, and blue line indicates a p-value = 0.001). Description of the genes is in Table 3 and detailed GEMMA results are given in Table S13.