Impact of Imputation, Reference Population Structure, and Single Nucleotide Polymorphism Panel Density on Accuracy of Genomic Evaluation in Purebred and Crossbred Populations

The objective of this study was to compare the accuracy of genomic breeding values prediction with different marker densities before and after the imputation in the simulated purebred and crossbred populations based on different scenarios of reference population and methods of marker effects estimation. The simulated populations included two purebred populations (lines A and B) and two crossbred populations (Cross and Backcross). Three different scenarios on selection of animals in the reference set including: (1) A high relationship with validation population, (2) Random, and (3) High inbreeding rate, were evaluated for imputation of validation population with the densities of 5 and 50K to 777K single marker polymorphism. Then, the accuracy of breeding values estimation in the validation population before and after the imputation was calculated by ABLUP, GBLUP, and SSGBLUP methods in two heritability levels of 0.25 and 0.5. The results showed that the highest accuracy of breeding values prediction in the purebred populations was obtained by GBLUP method and in the scenario of related reference population with validation set. However, in the crossbred population for the trait with low heritability (h(2) = 0.25), the highest accuracy of breeding values prediction in the weighting mechanism was equal to (lambda=0.2). Also, results showed that in the scenario of related reference population selection when 50K panel was used for genotype imputation to 777K SNPs, the prediction accuracy of genomic breeding values increased. But, in most scenarios of random and inbred reference set selection, there was no significant difference in the accuracy of genomic breeding values prediction between 5K and 50K SNPs after genotype imputation to 777K.

Automated summary

The study compared the accuracy of genomic breeding values prediction in purebred and crossbred populations using different marker densities and imputation scenarios. Results showed that GBLUP method performed best in purebred populations with related reference population, while weighting mechanism led to highest accuracy in crossbred populations with low heritability traits.