Copy number variants and selective sweeps in natural populations of the house mouse (Mus musculus domesticus)

Abstract

Copy-number variants (CNVs) may play an important role in early adaptations, potentially facilitating rapid divergence of populations. We describe an approach to study this question by investigating CNVs present in natural populations of mice in the early stages of divergence and their involvement in selective sweeps. We have analyzed individuals from two recently diverged natural populations of the house mouse (Mus musculus domesticus) from Germany and France using custom, high-density, comparative genome hybridization arrays (CGH) that covered almost 164 Mb and 2444 genes. One thousand eight hundred and sixty one of those genes we previously identified as differentially expressed between these populations, while the expression of the remaining genes was invariant. In total, we identified 1868 CNVs across all 10 samples, 200 bp to 600 kb in size and affecting 424 genic regions. Roughly two thirds of all CNVs found were deletions. We found no enrichment of CNVs among the differentially expressed genes between the populations compared to the invariant ones, nor any meaningful correlation between CNVs and gene expression changes. Among the CNV genes, we found cellular component gene ontology categories of the synapse overrepresented among all the 2444 genes tested. To investigate potential adaptive significance of the CNV regions, we selected six that showed large differences in frequency of CNVs between the two populations and analyzed variation in at least two microsatellites surrounding the loci in a sample of 46 unrelated animals from the same populations collected in field trappings. We identified two loci with large differences in microsatellite heterozygosity (Sfi1 and Glo1/Dnahc8 regions) and one locus with low variation across the populations (Cmah), thus suggesting that these genomic regions might have recently undergone selective sweeps. Interestingly, the Glo1 CNV has previously been implicated in anxiety-like behavior in mice, suggesting a differential evolution of a behavioral trait.

Keywords: CGH microarray; CNV; mice; natural populations; selective sweeps.

Figure 1

Overview of the experimental design.

Figure 2

Distribution of CNV length, number of probes and probe density in CNVs. (A) Comparison of distributions of probe densities in all samples and in each population separately and (B) Comparison of distributions of CNV size and number of probes per CNV in all samples and in each population separately. P-values from two-sample Kolmogorov-Smirnov test.

Figure 3

Comparison of distributions of CNV length, number of probes and probe density in differentiating (“diff”) vs. non-differentiating (“non-diff”) CNVs. P-values from two-sample Kolmogorov-Smirnov test.

Figure 4

Graphical overview of the Canonical Pathway analysis. This image was generated by Canonical Pathways analysis and shows genes that belong to the category of neuropathic pain signaling in dorsal horn neurons. Nodes in gray and purple depict genes from the CGH array submitted for the analysis: in gray, genes without CNVs and in purple, genes with CNVs. Solid lines around a node depict a kinase and dashed lines depict an ion transporter. Double lines indicate complex of several subunits.