RESOURCE DEVELOPMENT FACILITATING BOVINE GENOME SEQUENCE USE TO IMPROVE CATTLE PRODUCTION EFFICIENCY, PRODUCT QUALITY & ENVIRONMENTAL IMPACT
Location: Genetics, Breeding, & Animal Health
Title: Genomic Heritability of Beef Cattle Growth
Submitted to: Journal of Animal Science Supplement
Publication Type: Abstract Only
Publication Acceptance Date: February 23, 2009
Publication Date: July 12, 2009
Citation: Snelling, W.M., Kuehn, L.A., Thallman, R.M., Keele, J.W., Bennett, G.L. 2009. Genomic Heritability of Beef Cattle Growth [abstract]. Journal of Animal Science Supplement. 87(E-Suppl. 2):396. Abstract # W37.
Calf weights were examined to determine association between high-density SNP genotypes and growth, in order to estimate additive genetic variation explained by SNP. Data taken from Cycle VII of the U.S. Meat Animal Research Center Germplasm Evaluation Project included birth weight (BWT), 205-d adjusted preweaning gain (AWG), and 160-d postweaning gain (PWG) records of over 2,500 animals genotyped with the BovineSNP50 BeadChip. Polygenic and genomic direct and maternal variances were estimated in single-trait analyses. Fixed effects included sex, age of dam, year-season-location contemporary group, and covariates for calf and dam breed composition and heterosis. Direct and maternal additive polygenic and genomic effects, and maternal permanent environment effects were considered random. Polygenic effects were correlated according to a pedigree-based relationship matrix (A). Genomic effects were correlated with genotype-based relationship matrices (As) constructed from sets of SNP meeting single-SNP association criteria. The 4,242 animals represented in both A and As included 2,918 genotyped individuals. Genotypes of the remainder (mostly dams) were predicted by a single-locus BLUP procedure. SNP with minor allele frequencies < 0.05 were excluded. Tests to choose SNP, in order of set size, were Bonferroni-corrected P (Pb) < 0.05, false discovery rate (FDR) < 0.05, nominal P (Pn) < 0.01 and < 0.05, FDR > 0.90 (indicating SNP unlikely to affect phenotype), and all SNP. Unweighted As, with each SNP making equal contribution to the relationships, and weighted As, with contributions weighted according to estimated effects, were considered. Additive variance was split between polygenic and genomic components with Pb < 0.05 and FDR < 0.05, and shifted almost wholly to genomic components using Pn < 0.01 or < 0.05, and all SNP. Estimates of genomic variance were negligible for FDR > 0.90. Analyses with weighted As were more likely than with the corresponding unweighted As. The most likely model for each trait was with weighted As using Pn < 0.05, which yielded genomic direct heritability estimates of 0.46, 0.27 and 0.44 for BWT, AWG and PWG, and maternal heritability estimates of 0.29, 0.41 and 0.21.