Investigating genetic variants and pedigree-based diversity in Brazil's Friesian horse population

Abstract

The Friesian horse, recognized for its striking appearance, was established as a breed studbook in 1879 and remains popular in media. Despite its allure, the breed suffers from several genetic pathologies, including hydrocephaly (B3GALNT2, c.1423C > T), dwarfism (B4GALT7, c.50G > A) and distichiasis (ECA13:g.178714_195130del), alongside conditions of unknown origin such as megaesophagus, retained placenta, insect bite hypersensitivity, aortic rupture, and chronic idiopathic anhidrosis. The breed underwent a significant bottleneck early in its development, and 2 decades ago (2005) had an effective number of ancestors (Nae) of 16, around the same time as the first importations from the Netherlands to Brazil occurred. As of 2025, the Friesian Studbook in Brazil has fewer than 500 registered horses, raising concerns about a secondary geographic bottleneck and undesired genetic trait frequencies in this subpopulation. To assess potential pedigree-based detrimental effects, we analyzed data from 366 Friesians registered in Brazil, tracing pedigrees to their earliest recorded ancestors using global databases with the PurgeR package. Commercial genetic testing was conducted on 12 individuals actively-breeding, unrelated by parents and grandparents (3.27% of the total population) for known pathogenic, color and behavior variants of interest. Pedigree analysis indicated a reduced Nae of 11.93—lower than previous global population estimates—and an effective population size (Ne) of 33.03, with pedigree-based inbreeding coefficients (F_ped) ranging from 5.84% on individuals with equivalent complete generations (t) of 6.1, to 35.14% with t of 11.7. Genetic testing resulted in a 33% carrier rate for dwarfism and 16.6% for hydrocephaly. We also identified 2 individuals as HPS5^EDXW1/EDXW3, and one individual as KIT^W20/+, variants associated with white spotting patterns in other horse breeds, despite no phenotypic expression of white in the individuals. Additionally, we observed allelic heterogeneity in the DRD4 temperament locus, where DRD4^G/G = 75%, DRD4^G/A = 16.7% and DRD4^A/A = 8.3%, the latter associated with a higher vigilance phenotype. The pedigree-based probability of homozygosity for recessive alleles on a locus being identical by descent in this population is estimated as 10.7% (SD ± 3%), reiterating the importance of genetic testing for selection. These findings underscore the critical need for genetic testing and population monitoring to preserve genetic diversity and mitigate health risks in the Friesian population, particularly within isolated geographic subgroups.