Evidence of dynamic selective pressures on VR repertoires in both laboratory- and wild-derived strains. (A) The ratio of non-synonymous to synonymous SNPs in VRs is similar between wild- and lab-derived strains (mean + SEM. Two-tailed t-test, P = 0.622), but higher than the genome averages (dashed lines: lab-derived = 0.54, wild-derived = 0.49 calculated from the data in ). (B) A schematic representation of the proposed secondary structure of a V2R receptor protein. The N-terminal domain (blue) is likely to be extracellular allowing it to interact with its pheromone ligand. (C) The normalised distribution of SNPs reveals significantly more functional variation is found in the N-terminal domains of V2Rs than trans-membrane domains (Two-way ANOVA, variance in domain F1,14 = 138, P < 0.0001, variance in strain F1,14 = 1.5, P = 0.24. Bonferroni post hoc test, *** = P < 0.001). (D) The distribution of termination codons across VRs. The VR domains are indicated by shading (N- and C-termini in dark grey, trans-membrane domains in lighter greys). The relative positions of the termination codon found in each gene are indicated beneath. The strain each termination codon is found in is indicated by colour and shape (lab-derived strains by circles, wild-derived strained by triangles), above. Note some termination codons are found in multiple strains and some genes have multiple termination codons (e.g. Vmn2r120). (E) The proportion of genes analysed with evidence of truncation, frame shift, deletion or duplication, suggesting the functional VR repertoire is highly variable between strains of mice. Any gene that contained more than one of these (for example, was truncated and had a deletion) was counted only once in the following order: duplication, deletion, truncation, frame shift.