Use of genetic variance to determine weighting factors for genomic information used in single step genomic evaluation

Abstract

Single step genomic BLUP (SS-GBLUP) has recently been implemented in the Australian beef cattle and sheep industries. Selection of the weighting factor (λ) to combine the genomic and pedigree relationship matrixes could affect the accuracies of EBV. Cross validation has been used as a method to determine the optimal λ. This study estimated genetic variances using models fitting the genomic relationship matrix (G, Vg), the pedigree relationship matrix (A, Va), or both (Vt), and used the ratio Vg/Vt as an alternative approach to identify the optimal λ. Results showed that the ratio of genetic variances (Vg/Vt) aligned reasonably well with the optimal values of λ for most traits. Compared with cross validation, this approach was computationally fast, and can be used as an alternative to identify the optimal values of λ in SS-GBLUP.