Zhu Shaohua, Zhao Hongchang, Han Mei, Yuan Chao, Guo Tingting, Liu Jianbin, Yue Yaojing, Qiao Guoyan, Wang Tianxiang, Li Fanwen, Gun Shuangbao, Yang Bohui
College of Animal Science and Technology, Gansu Agricultural University, Lanzhou, China.
Animal Science Department, Lanzhou Institute of Husbandry and Pharmaceutical Sciences, Chinese Academy of Agricultural Sciences, Lanzhou, China.
Front Vet Sci. 2020 Nov 11;7:573692. doi: 10.3389/fvets.2020.573692. eCollection 2020.
Dominant genetic effects may provide a critical contribution to the total genetic variation of quantitative and complex traits. However, investigations of genome-wide markers to study the genomic prediction (GP) and genetic mechanisms of complex traits generally ignore dominant genetic effects. The increasing availability of genomic datasets and the potential benefits of the inclusion of non-additive genetic effects in GP have recently renewed attention to incorporation of these effects in genomic prediction models. In the present study, data from 498 genotyped Alpine Merino sheep were adopted to estimate the additive and dominant genetic effects of 9 wool and blood traits via two linear models: (1) an additive effect model (MAG) and (2) a model that included both additive and dominant genetic effects (MADG). Moreover, a method of 5-fold cross validation was used to evaluate the capability of GP in the two different models. The results of variance component estimates for each trait suggested that for fleece extension rate (73%), red blood cell count (28%), and hematocrit (25%), a large component of phenotypic variation was explained by dominant genetic effects. The results of cross validation demonstrated that the MADG model, comprising additive and dominant genetic effects, did not display an apparent advantage over the MAG model that included only additive genetic effects, i.e., the model that included dominant genetic effects did not improve the capability for prediction of the genomic model. Consequently, inclusion of dominant effects in the GP model may not be beneficial for wool and blood traits in the population of Alpine Merino sheep.
显性遗传效应可能对数量性状和复杂性状的总遗传变异做出关键贡献。然而，利用全基因组标记来研究复杂性状的基因组预测（GP）和遗传机制的研究通常忽略了显性遗传效应。随着基因组数据集的日益丰富，以及在GP中纳入非加性遗传效应的潜在益处，最近人们重新关注在基因组预测模型中纳入这些效应。在本研究中，采用来自498只基因分型的阿尔卑斯美利奴羊的数据，通过两个线性模型估计9个羊毛和血液性状的加性和显性遗传效应：（1）加性效应模型（MAG）和（2）一个同时包含加性和显性遗传效应的模型（MADG）。此外，采用5折交叉验证方法评估两种不同模型中GP的能力。每个性状的方差成分估计结果表明，对于羊毛延伸率（73%）、红细胞计数（28%）和血细胞比容（25%），显性遗传效应解释了很大一部分表型变异。交叉验证结果表明，包含加性和显性遗传效应的MADG模型相对于仅包含加性遗传效应的MAG模型没有表现出明显优势，即包含显性遗传效应的模型并没有提高基因组模型的预测能力。因此，在GP模型中纳入显性效应可能对阿尔卑斯美利奴羊群体的羊毛和血液性状没有益处。
