High-resolution population structure and runs of homozygosity reveal the genetic architecture of complex traits in the Lipizzan horse.

BACKGROUND: The sample ascertainment bias due to complex population structures remains a major challenge in genome-wide investigations of complex traits. In this study we derived the high-resolution population structure and levels of autozygosity of 377 Lipizzan horses originating from five different European stud farms utilizing the SNP genotype information of the high density 700 k Affymetrix Axiom™ Equine genotyping array. Scanning the genome for overlapping runs of homozygosity (ROH) shared by more than 50% of horses, we identified homozygous regions (ROH islands) in order to investigate the gene content of those candidate regions by gene ontology and enrichment analyses. RESULTS: The high-resolution population network approach revealed well-defined substructures according to the origin of the horses (Austria, Slovakia, Croatia and Hungary). The highest mean genome coverage of ROH (SROH) was identified in the Austrian (SROH = 342.9), followed by Croatian (SROH = 214.7), Slovakian (SROH = 205.1) and Hungarian (SROH = 171.5) subpopulations. ROH island analysis revealed five common islands on ECA11 and ECA14, hereby confirming a closer genetic relationship between the Hungarian and Croatian as well as between the Austrian and Slovakian samples. Private islands were detected for the Hungarian and the Austrian Lipizzan subpopulations. All subpopulations shared a homozygous region on ECA11, nearly identical in position and length containing among other genes the homeobox-B cluster, which was also significantly (p < 0.001) highlighted by enrichment analysis. Gene ontology terms were mostly related to biological processes involved in embryonic morphogenesis and anterior/posterior specification. Around the STX17 gene (causative for greying), we identified a ROH island harbouring the genes NR4A3, STX17, ERP44 and INVS. Within further islands on ECA14, ECA16 and ECA20 we detected the genes SPRY4, NDFIP1, IMPDH2, HSP90AB1, whereas SPRY4 and HSP90AB1 are involved in melanoma metastasis and survival rate of melanoma patients in humans. CONCLUSIONS: We demonstrated that the assessment of high-resolution population structures within one single breed supports the downstream genetic analyses (e.g. the identification of ROH islands). By means of ROH island analyses, we identified the genes SPRY4, NDFIP1, IMPDH2, HSP90AB1, which might play an important role for further studies on equine melanoma. Furthermore, our results highlighted the impact of the homeobox-A and B cluster involved in morphogenesis of Lipizzan horses.

Genome-Wide Homozygosity Patterns and Evidence for Selection in a Set of European and Near Eastern Horse Breeds.

Intensive artificial and natural selection have shaped substantial variation among European horse breeds. Whereas most equine selection signature studies employ divergent genetic population structures in order to derive specific inter-breed targets of selection, we screened a total of 1476 horses originating from 12 breeds for the loss of genetic diversity by runs of homozygosity (ROH) utilizing a 670,000 single nucleotide polymorphism (SNP) genotyping array. Overlapping homozygous regions (ROH islands) indicating signatures of selection were identified by breed and similarities/dissimilarities between populations were evaluated. In the entire dataset, 180 ROH islands were identified, whilst 100 islands were breed specific, all other overlapped in 36 genomic regions with at least one ROH island of another breed. Furthermore, two ROH hot spots were determined at horse chromosome 3 (ECA3) and ECA11. Besides the confirmation of previously documented target genes involved in selection for coat color (MC1R, STX17, ASIP), body size (LCORL/NCAPG, ZFAT, LASP1, HMGA2), racing ability (PPARGC1A), behavioral traits (GRIN2B, NTM/OPCML) and gait patterns (DMRT3), several putative target genes related to embryonic morphogenesis (HOXB), energy metabolism (IGFBP-1, IGFBP-3), hair follicle morphogenesis (KRT25, KRT27, INTU) and autophagy (RALB) were highlighted. Furthermore, genes were pinpointed which might be involved in environmental adaptation of specific habitats (UVSSA, STXBP4, COX11, HLF, MMD).