Construction of realized genomic relationships matrix for BLUP of breeding values
Linear mixed models that utilize realized genomic relationships matrices could predict breeding values more accurately than those that use expected average genetic covariances derived from pedigrees. An approach that can be explored is the solving of mixed model equations by using the inverse of the genomic relationship matrix (G), in place of the A matrix. This may allow better estimation of individual allele effects, followed by summation across loci to obtain genomic estimated breeding values in a marker-only model. The generation of the G matrix based on a set of biallelic SNP markers genotyped for a clonal population of loblolly pine and its application into calculation of accurate genomic estimated breeding values is discussed. The correlation between the genomic estimated breeding values and breeding values based on a pedigree-based model is used to determine the potential gain using markers in a forest tree breeding scenario.
Download this file:
Download this file — PDF document, 38KbDetails
Author(s): Jaime Zapata-Valenzuela