Genome-wide architecture of reproductive isolation in a naturally occurring hybrid zone between Mus musculus musculus and M. m. domesticus

Abstract

Studies of a hybrid zone between two house mouse subspecies (Mus musculus musculus and M. m. domesticus) along with studies using laboratory crosses reveal a large role for the X chromosome and multiple autosomal regions in reproductive isolation as a consequence of disrupted epistasis in hybrids. One limitation of previous work has been that most of the identified genomic regions have been large. The goal here is to detect and characterize precise genomic regions underlying reproductive isolation. We surveyed 1401 markers evenly spaced across the genome in 679 mice collected from two different transects. Comparisons between transects provide a means for identifying common patterns that likely reflect intrinsic incompatibilities. We used a genomic cline approach to identify patterns that correspond to epistasis. From both transects, we identified contiguous regions on the X chromosome in which markers were inferred to be involved in epistatic interactions. We then searched for autosomal regions showing the same patterns and found they constitute about 5% of autosomal markers. We discovered substantial overlap between these candidate regions underlying reproductive isolation and QTL for hybrid sterility identified in laboratory crosses. Analysis of gene content in these regions suggests a key role for several mechanisms, including the regulation of transcription, sexual conflict and sexual selection operating at both the postmating prezygotic and postzygotic stages of reproductive isolation. Taken together, these results indicate that speciation in two recently diverged (c. 0.5 Ma) house mouse subspecies is complex, involving many genes dispersed throughout the genome and associated with distinct functions.

Fig. 1

Genotype probabilities given hybrid index (genomic clines) for 85 X-linked markers as provided by the Introgress program (Gompert & Buerkle 2009) for (a) BV and (b) CZ transect (solid lines: homozygotes for M. m. domesticus alleles; dashed lines: heterozygotes). Null expectations (light grey: homozygotes for M. m. domesticus alleles; dark grey: heterozygotes) are based on the permutation of autosomal data (1316 markers). Hybrid index used from Wang et al. (2011). Red lines represent average genomic clines for X-linked markers.

Fig. 2

Comparison of introgression models for pairs of markers plotted along their physical positions on the X chromosome (BV: above diagonal; CZ: below diagonal). (a) log-likelihood ratios (ΔLLs) for pairs of markers. The more similar behaviour of the two markers the lower ΔLL. (b) statistical significance of pairwise comparisons (red: behaviour of the two markers is not significantly different). Plotted in the R statistical environment (http://www.r-project.org/) using modified version of LDHEATMAP package (Shin et al. 2006).

Fig. 3

The X-linked uniform regions (XURs) were defined based on the P-value estimates for the pairwise comparisons of X-linked markers in each transect separately (black lines). Red lines signify regions containing markers not significantly different from neutral expectations based on the permutation of autosomal data (1316 markers). The overlapping regions between XURs are considered as shared X-linked uniform regions (SXURs) (blue rectangles).

Fig. 4

Genomic clines for 29 X-linked markers from within the 10 SXURs plotted for each transect separately. Probability of genotype (M. m. domesticus homozygotes—solid line; heterozygotes—dashed line) plotted on hybrid index. Null expectations (light grey: homozygotes for M. m. domesticus alleles; dark grey: heterozygotes) are based on the permutation of autosomal data (1316 markers). Hybrid index used from Wang et al. (2011). Numbers correspond to the SXURs numbers in Fig. 3.

Fig. 5

Genome location of autosomal markers. Grey ticks depict all of the 1316 autosomal markers. Blue and red circles depict candidate epistatically interacting markers in BV and CZ transect, respectively (P-value_pairwise > 0.001 and P-value_introgress ≤ 0.001).

Fig. 6

Candidate autosomal regions associated with reproductive isolation in the house mouse. Shared regions between the two transects exhibiting epistasis (this study) are shown in red. The QTL regions associated with hybrid sterility phenotypes (White et al. 2011) are represented as blue lines. Phenotype abbreviations, TW, relative testis weight; SD, sperm density; PC1, sperm head morphology; STA, seminiferous tubule area; DBT, distal bent tail; TAS, total abnormal sperm; H/T, headless/tailless correspond to those in White et al. (2011). Genes identified in Dean et al. (2008, 2011) are coloured in violet and those identified by Dorus et al. (2010) are coloured in green.