GVCBLUP: A Computing Package for Genomic Prediction and REML Estimation of Variance Components for Additive and Dominance Effects using SNP Markers

The GVCBLUP package currently contains 25 computer programs for genomic prediction (GBLUP) and REML estimation (GREML) for research and education. These programs considered two marker-individual configurations of ‘q<m’ and ‘q>m’, three definitions of genomic additive and dominance relationship matrices and two mixed model implementations (CE and MME implementations), where q = number of individuals, m = number of markers, CE = the conditional expectation of breeding values or dominance deviations given phenotypic observations with fixed effects estimated by BLUE, and MME = mixed model equations. The three definitions of genomic relationships include additive and dominance relationship matrices using the total heterozygosity or the average diagonal elements in the denominator of each relationship matrix, and additive and dominance correlation matrices with diagonal elements of ‘1’ and off-diagonal elements in the range of -1 and 1. Of the 25 programs, six programs are for GREML of additive and dominance variance components and associated heritabilities, six programs for GREML of additive variance component only, six programs for GBLUP of additive and dominance effects, six programs for GBLUP of additive effects only, and one program for calculating additive and dominance relationship and correlation matrices for the three definitions of such matrices. The CE implementation applies to both cases of ‘q<m’ and ‘q>m’ whereas the MME implementation currently is available for ‘q>m’ only. Majority of these programs have been evaluated using simulation data with 1K causal variants, and 1K, 3K, 7K and 41K SNP markers.