Haplotype Discovery and an Imputation Resource for the Domestic Horse

The development of a low-density chip for the horse has enabled successful genetic
studies examining domestication and selection, disease and performance trait
mapping and population dynamics. However, many horse breeds have extensive
population structure and low linkage disequilibrium which severely limit conventional
mapping approaches with the current SNP chip. To address these issues, we will
leverage whole genome sequencing data generated by the equine genome
community. To date we have processed Illumina HiSeq data from 163 domestic
horses representing 32 distinct breeds. Over 57 billion reads have been produced
from these individuals with a mean depth of coverage of 10X. We have implemented
a large scale mapping and variant calling pipeline to improve sensitivity and
specificity of variant calling by identifying variants using both GATK and samtools;
calling variants across individuals from related breeds; and identifying a set of high
quality variants which can be used for variant recalibration. For example, for 6
Morgan horses at 12X coverage, 12M and 11M variants were discovered by GATK
and samtools and 10M were called by both. We are using this data to design 2M and
670K genotyping arrays; determine fine scale haplotype structure and create a
haplotype map; and develop an imputation pipeline which will enable those with
current generation genotyping platforms to increase the number of markers in their
samples with no additional overhead. Summary data from mapping and calling
pipelines and a selection of variants for the genotyping arrays, initial haplotype
analysis and imputation pipelines will be presented.