PAG XXVI - HapDab: A haplotype map and imputation resource in the horse

The recently developed Equine genotyping array (MNEc670k) was designed for genotype imputation. Variants (SNPs) on the array were chosen to differentiate among patterns of linkage disequilibrium (LD) between diverse horse populations. This enables a smaller set of SNPs (~670k) to be imputed to a higher density (~2M) using haplotypes characterized in a reference population. Moreover, LD decays at different rates throughout the genome as well as between different breeds resulting in differing haplotype lengths. This is important for association studies where markers are unlikely to be functional, instead being in LD with the causal variant.
Using a reference population of 485 horses representing 15 breeds, we calculated breed specific haplotypes throughout the genome around each of 2M core SNPs. We cataloged these haplotypes along with LD decay in a relational database to enable data-driven SNP-to-gene mapping for association studies. Using this same reference population, we’ve implemented an equine imputation resource. Input samples genotyped on the MNEc670k array (or previous generation arrays) are phased and phylogenetically clustered in order to determine an imputation population.