Accuracy of Imputation to High Density SNP Data in Multibreed Beef Cattle

Imputation enables the determination of missing SNP genotypes in animals genotyped with a low-density panel by using information from a reference population genotyped with a higher density panel. Imputation enables more genotypes at higher density for genomic selection programs. In this study, population imputation, utilizing linkage disequilibrium among markers, was implemented using the software BEAGLE and FIMPUTE 2.2 in a multi-breed, crossbred taurine beef cattle population genotyped (n = 96) with both the Illumina High-Density Bovine BeadChip Array (BOVINEHD 777,962 SNP) and Affymetrix Axiom Genome-Wide BOS1 Array (BOS1 648,874 SNP). Multiple scenarios combining 80 animals in the reference population were averaged for imputation accuracy and average results from Fimpute are presented. Imputation accuracy for a 6K panel to BOVINEHD and BOS1 were 78.65 and 83.27%, respectively. Imputation accuracy from 50K to BOVINEHD and BOS1 were 91.74 and 90.95%, respectively. Accuracy of 6K, 50K, BOVINEHD and BOS1 to a combined marker panel (1,261,128 SNP) was 79.53, 91.33, 98.26 and 98.03%, respectively. Higher imputation accuracies were obtained when the breed composition of imputed animals was well represented in the reference population. FImpute outperformed Beagle in all scenarios with an average increase in accuracy of 2.67% while reducing run-time by 20 to 48 times. Relatively high imputation accuracy for the combined HD panel was possible from either HD panel with a relatively small reference population.