Genotype Imputation of Crossbred Chickens

Breeders could further improve the robustness of commercial chickens, which are typically three-way or four-way crosses, if they can select parental pure lines based on crossbred performance under field conditions. However, it is difficult and expensive to collect pedigree data (required in conventional selection) from commercial chickens. Unlike convectional selection, genomic selection is immune to this problem: using thousands of single nucleotide polymorphisms (SNPs), breeders can establish genomic relationship between commercial crossbreds and their collaterals in pure lines (i.e. selection candidates). Genomic selection, however, has not been extensively implemented in chickens mainly because of high costs of genotyping a large number of SNPs in a large number of animals. While the costs of genomic selection could be lowered using low-density (LD) SNP panels with methods that impute missing genotypes, there is limited information on usefulness of these approaches in crossbred animals. We designed a 6K LD panel based on the information from Illumina 60K SNP genotypes of three parental pure lines. Then we used the LD panel to impute 60K SNP genotypes of crossbreds using the pure lines, crossbreds or a 50:50 mix of pure lines and crossbreds as reference populations. Accuracy of imputation was over 96% for most scenarios, indicating that using an LD panel, breeders can impute—with high accuracies—high-density genotypes of crossbred chickens.