Moghaddar N, van der Werf J H J
School of Environmental and Rural Science, University of New England, Armidale, NSW, Australia.
Cooperative Research Centre for Sheep Industry Innovation, Armidale, NSW, Australia.
J Anim Breed Genet. 2017 Dec;134(6):453-462. doi: 10.1111/jbg.12287. Epub 2017 Aug 17.
The objectives of this study were to estimate the additive and dominance variance component of several weight and ultrasound scanned body composition traits in purebred and combined cross-bred sheep populations based on single nucleotide polymorphism (SNP) marker genotypes and then to investigate the effect of fitting additive and dominance effects on accuracy of genomic evaluation. Additive and dominance variance components were estimated in a mixed model equation based on "average information restricted maximum likelihood" using additive and dominance (co)variances between animals calculated from 48,599 SNP marker genotypes. Genomic prediction was based on genomic best linear unbiased prediction (GBLUP), and the accuracy of prediction was assessed based on a random 10-fold cross-validation. Across different weight and scanned body composition traits, dominance variance ranged from 0.0% to 7.3% of the phenotypic variance in the purebred population and from 7.1% to 19.2% in the combined cross-bred population. In the combined cross-bred population, the range of dominance variance decreased to 3.1% and 9.9% after accounting for heterosis effects. Accounting for dominance effects significantly improved the likelihood of the fitting model in the combined cross-bred population. This study showed a substantial dominance genetic variance for weight and ultrasound scanned body composition traits particularly in cross-bred population; however, improvement in the accuracy of genomic breeding values was small and statistically not significant. Dominance variance estimates in combined cross-bred population could be overestimated if heterosis is not fitted in the model.
本研究的目的是基于单核苷酸多态性（SNP）标记基因型，估计纯种和杂交绵羊群体中几个体重和超声扫描身体组成性状的加性和显性方差分量，然后研究拟合加性和显性效应对基因组评估准确性的影响。使用从48,599个SNP标记基因型计算出的动物间加性和显性（协）方差，在基于“平均信息约束最大似然法”的混合模型方程中估计加性和显性方差分量。基因组预测基于基因组最佳线性无偏预测（GBLUP），并基于随机10倍交叉验证评估预测准确性。在不同的体重和扫描身体组成性状中，纯种群体的显性方差占表型方差的比例为0.0%至7.3%，杂交群体为7.1%至19.2%。在杂交群体中，考虑杂种优势效应后，显性方差范围降至3.1%和9.9%。考虑显性效应显著提高了杂交群体中拟合模型的似然性。本研究表明，体重和超声扫描身体组成性状存在显著的显性遗传方差，尤其是在杂交群体中；然而，基因组育种值准确性的提高很小且在统计学上不显著。如果模型中未拟合杂种优势，杂交群体中的显性方差估计可能会被高估。
