Selection in Australian Thoroughbred horses acts on a locus associated with early two-year old speed

Thoroughbred horse racing is a global sport with major hubs in Europe, North America, Australasia and Japan. Regional preferences for certain traits have resulted in phenotypic variation that may result from adaptation to the local racing ecosystem. Here, we test the hypothesis that genes selected for regional phenotypic variation may be identified by analysis of selection signatures in pan-genomic SNP genotype data. Comparing Australian to non-Australian Thoroughbred horses (n = 99), the most highly differentiated loci in a composite selection signals (CSS) analysis were on ECA6 (34.75-34.85 Mb), ECA14 (33.2-33.52 Mb and 35.52-36.94 Mb) and ECA16 (24.28-26.52 Mb) in regions containing candidate genes for exercise adaptations including cardiac function (ARHGAP26, HBEGF, SRA1), synapse development and locomotion (APBB3, ATXN7, CLSTN3), stress response (NR3C1) and the skeletal muscle response to exercise (ARHGAP26, NDUFA2). In a genome-wide association study for field-measured speed in two-year-olds (n = 179) SNPs contained within the single association peak (33.2-35.6 Mb) overlapped with the ECA14 CSS signals and spanned a protocadherin gene cluster. Association tests using higher density SNP genotypes across the ECA14 locus identified a SNP within the PCDHGC5 gene associated with elite racing performance (n = 922). These results indicate that there may be differential selection for racing performance under racing and management conditions that are specific to certain geographic racing regions. In Australia breeders have principally selected horses for favourable genetic variants at loci containing genes that modulate behaviour, locomotion and skeletal muscle physiology that together appear to be contributing to early two-year-old speed.