Fine-Scale Recombination Map for the Domestic Horse

Recombination events have been shown in eukaryotes to cluster along chromosomes in small regions called ‘hotspots’. In humans and mice, total genetic map length and local recombination landscapes vary by gender as well as across populations, suggesting that selective pressures play a role in recombination rate distribution across the genome. To create a gender- and breed-averaged recombination map for the horse, 513 horses genotyped for ~1.8 million SNPs on the MNEq2M array were divided into 20 groups of 20 random individuals to estimate population recombination rates (4N_er) using the rhomap program included in the LDhat package. Chromosomes were split into intervals of 2000 SNPs with a 200-SNP overlap between adjacent windows, and each analysis was run for 10 million iterations with a 100,000 burn-in period. To create breed-specific maps this procedure was performed in breeds in which 20 or more individuals had been genotyped. These rate estimates were used to predict recombination hotspots both across breeds and within individual breeds using the LDhot program. To convert recombination rates to centimorgans, the effective population size (N_e) of each population was calculated through approximate Bayesian computation using 1 million simulated datasets with known contemporary N_e. Our data show clear differences in local recombination rates across breeds and will therefore be useful in improving imputation accuracy between genotyping arrays. We also seek to identify sequence motifs in hotspot regions as well as regions with low recombination to gain insight into the control of recombination initiation in the horse.