Limited Evidence for Parallel Evolution Among Desert-Adapted Peromyscus Deer Mice

Abstract

Warming climate and increasing desertification urge the identification of genes involved in heat and dehydration tolerance to better inform and target biodiversity conservation efforts. Comparisons among extant desert-adapted species can highlight parallel or convergent patterns of genome evolution through the identification of shared signatures of selection. We generate a chromosome-level genome assembly for the canyon mouse (Peromyscus crinitus) and test for a signature of parallel evolution by comparing signatures of selective sweeps across population-level genomic resequencing data from another congeneric desert specialist (Peromyscus eremicus) and a widely distributed habitat generalist (Peromyscus maniculatus), that may be locally adapted to arid conditions. We identify few shared candidate loci involved in desert adaptation and do not find support for a shared pattern of parallel evolution. Instead, we hypothesize divergent molecular mechanisms of desert adaptation among deer mice, potentially tied to species-specific historical demography, which may limit or enhance adaptation. We identify a number of candidate loci experiencing selective sweeps in the P. crinitus genome that are implicated in osmoregulation (Trypsin, Prostasin) and metabolic tuning (Kallikrein, eIF2-alpha kinase GCN2, APPL1/2), which may be important for accommodating hot and dry environmental conditions.

Keywords: Peromyscus; dehydration; desert; parallel evolution; thermoregulation.

Figure 1.

Geographic ranges of the 3 Peromyscus species examined in this study. Major North American deserts are denoted by diagonal hashing. Peromyscus crinitus range is in red/medium grey, Peromyscus eremicus in blue/dark grey, and Peromyscus maniculatus in yellow/light grey. Areas of sympatry are denoted by additive color overlap. Collection localities are labeled with white dots and include the Motte Rimrock Reserve (MOT), Elliot Chaparral Reserve (ELL), and Philip L. Boyd Deep Canyon Desert Research Center (DCDRC). See online version for full colors.

Figure 2.

Distributions of effective population size (N_e) through time for Peromyscus crinitus (red/top), Peromyscus eremicus (blue/bottom), and Peromyscus maniculatus (yellow/middle) based on a generation time of 6 months (0.5 years) and a general mammalian mutation rate of 2.2 × 10⁻⁹ substitutions/site/year. Note that the P. maniculatus genome was sequenced from a captive individual and therefore does not reflect natural populations’ trends of this species. CLR scores for P. crinitus based on Sweepfinder2 results. Values above the horizontal red line surpass the 99.9^th percentile. The top 5 or fewer unique genes are labeled for each chromosome. See online version for full colors.

Figure 3.

REVIGO plots of enriched functional groups for Peromyscus crinitus (top row) and Peromyscus eremicus (bottom row) based on functional annotation of the 2 nearest protein-coding genes to each site (dataset II) identified as the subject of a selective sweep. Darker colors indicate greater significance. MP, metabolic process, MB, membrane-bound. See online version for full colors.

Figure 4.

CLR scores for Peromyscus crinitus based on Sweepfinder2 results. Values above the horizontal dashed line surpass the 99.9^th percentile. The top 5 or fewer unique genes are labeled for each chromosome.

Figure 5.

Overlap in proximal gene names (top row) and enriched GO terms (bottom row) for datasets I (left column), II (center), and III (right). * indicates significant overlap between species. See online version for full colors.