Accuracy of genotype imputation with different low density panels in Braford and Hereford cattle.

The main objective of this research was to test alternative low density SNP panels to impute Illumina 50K SNP panel genotypes in Braford and Hereford cattle. Genotypes from 3,768 Hereford, Braford and Nellore animals were used for testing imputation from low density SNP panels (3K, 6K, 8K, 15K and 20K) to the Illumina 50K SNP panel, under four different scenarios: including or not Nellore genotypes in the reference population in combination with the use or not of pedigree information. There were no significant differences in imputation accuracy among these four scenarios within each panel. However, significant differences between panels were found. The best accuracy was given by a customized 15K SNP panel, with an overall genotype concordance rate of 0.977, with 93.3% of the animals imputed with a concordance rate above 0.95. The concordance rates for the other SNP panels were 0.872, 0.952, 0.957 and 0.958 for 3K, 6K, 8K and 20K SNP panel, respectively. Therefore, in the Braford/Hereford population considered in this study, all the alternative panels denser than 3K could be used for imputing to the 50K SNP panel with an overall high imputation accuracy. However, the best results were obtained with the customized 15K SNP instead of the alternative commercial panels. The use of Nellore sire genotypes and pedigree information did not increase accuracy of imputation in this population