Investigating the fluctuations of residual polygenic variance on predictive ability of genomic breeding values of crossbred progeny

Genomic evaluation of crossbred progeny due to their limited pedigree and lack of genotyping and performance accessibility is inevitably based on the information from purebred parental populations to increase crossbred performance. Despite the fact that single-step genomic evaluation (ssGBLUP) method can use information from both genotyped and non-genotyped animals simultaneously, leading to more accurate genomic estimated breeding values, but due to the incompatibility of the genomic and pedigree relationship matrices, it might lead to dispersion (inflation/deflation) and bias in the estimated breeding values higher than that with the BLUP method. Therefore, in this study, the prediction ability of the genomic breeding value of crossbred progeny was investigated based on the ssGBLUP method and simulation data of the purebred parents and crossbred population, considering the different scaling factor of residual polygenic variance. Based on the results of this study, the use of different ratio of residual polygenic (β) variance to incorporate the genomic and pedigree relationship matrices did not considerably affect the accuracy, bias and dispersion of the predicted genomic breeding values in crossbred progeny. In addition, the effect of proportion of residual polygenic variance (β) in blending the genomic and pedigree relationship matrices does not significantly affect the convergence point. Therefore, to simplify the model, the default value of β (0.05) might be used in the inverse of the relationship matrix.