Effect of founder breeds on genotype imputation accuracy in Canchim cattle

Genotype imputation is a technique used to infer unobserved genotypes based on reference panels, allowing increased marker density and cost-effective optimization for genomic selection. This study aimed to evaluate whether the inclusion of genotypes from the founder breeds Nelore (NE) and Charolais (CH) improves the imputation accuracy in the composite beef cattle breed Canchim (CA). The populations studied consisted of 804 NE, 897 CH, and 392 CA animals, all genotyped using high-density panels (777,962 SNP – single nucleotide polymorphisms). CA animals had their genotypes masked to simulate a medium-density panel (54,609 SNP). Fourteen imputation scenarios were evaluated, varying according to breed, sex, year of birth, and lineage. Imputation accuracy was determined based on the percentage of correctly imputed genotypes (PERC) and the squared Pearson’s correlation between observed and imputed genotypes (R2). PERC values ranged from 66.52% to 97.39% and R² from 0.6352 to 0.9780. The scenarios that included NE, CH, and CA (males or animals born before 2004) as the reference population for imputing CA females or CA animals born after 2004 showed the highest imputation accuracies. Therefore, the use of founder breeds in the reference population improves the accuracy of genotype imputation in CA cattle. The results indicate that a multibreed reference population, incorporating founder breeds, could provide a more robust and informative genetic basis for imputing composite cattle.