基于元共祖的基因组联合育种模拟研究  

Simulation Study on Joint Genomic Breeding Using Metafounders

作　　者：庞志旭 张洪志 乔利英[1] 王万年 潘洋洋[1] 刘文忠[1] PANG Zhixu;ZHANG Hongzhi;QIAO Liying;WANG Wannian;PAN Yangyang;LIU Wenzhong(College of Animal Science, Shanxi Agricultural University, Taigu 030801, China)

机构地区：[1]山西农业大学动物科学学院,太谷030801

出　　处：《畜牧兽医学报》2022年第7期2172-2181,共10页ACTA VETERINARIA ET ZOOTECHNICA SINICA

基　　金：国家自然科学基金(31972560);山西省攻关项目(011029);“雁云白羊”新品种(系)培育(YTGC126)。

摘　　要：旨在将整合元共祖的一步法(single-step genomic best linear unbiased prediction with metafounders,MF-SSGBLUP)应用到基因组联合育种中,并与其他经典基因组选择方法进行比较分析。本研究使用QMSim软件模拟3个系谱相互独立的奶牛群体;分别使用广义最小二乘法(generalized least squares,GLS)和原始方法(naive,NAI)估计不同群体间的祖先关系矩阵Γ;将MF-SSGBLUP、SSGBLUP和BLUP用于3个模拟群体的联合育种,评估各方法在遗传参数和育种值估计方面的差异。在不同遗传力下,GLS所得的Γ矩阵在对角线元素上略低于NAI法,在非对角线元素上没有明显差异,且基因组关系矩阵与基于元共祖构建的亲缘关系矩阵对角线元素相关系数(0.750~0.775)高于基因组关系矩阵与传统的亲缘关系矩阵相关系数(0.508~0.572)。MF-SSGBLUP遗传力估计值(0.138、0.140、0.297和0.298)与当代群体遗传力(0.107和0.296)的偏差小于其余两种方法(0.145、0.173、0.273和0.340),且MF-SSGBLUP估计育种值准确性(0.888~0.908)高于SSGBLUP法(0.863~0.876)和BLUP法(0.854~0.871)。表明,MF-SSGBLUP的遗传参数估计值无偏性更好,估计育种值准确性更高。根据上述模拟数据结果表明,在联合育种中,整合元共祖的基因组选择方法优于其他经典基因组选择方法。This study aimed to apply the single-step genomic best linear unbiased prediction with metafounders(MF-SSGBLUP) to joint genomic breeding and compare it with other classical genomic selection methods. QMSim software was used to simulate 3 dairy cattle populations with independent pedigrees;The generalized least squares(GLS) and naive(NAI) methods were used to estimate the ancestral relationship matrix Γ between different populations. MF-SSGBLUP, SSGBLUP and BLUP were used respectively to joint breeding for the simulated populations, and the performance of each method in estimating genetic parameters and breeding values was evaluated. For different heritabilities, the Γ matrix obtained by GLS was slightly lower than that obtained by NAI method in diagonal elements, but there was no significant difference in non-diagonal elements. The correlation coefficient in diagonal elements between the genomic and genetic relationship matrices based on metafounders(0.750-0.775) was higher than that between genomic and traditional relationship matrices(0.508-0.572). The deviations of heritability estimates by MF-SSGBLUP(0.138, 0.140, 0.297 and 0.298) from the current population heritability(0.107 and 0.296) were smaller than those of the other two methods(0.145, 0.173, 0.273 and 0.340). Correspondingly, the accuracies of estimated breeding values by MF-SSGBLUP(0.888-0.908) were higher than that by SSGBLUP(0.863-0.876) and BLUP(0.854-0.871). The results showed that MF-SSGBLUP had less biased estimates of genetic parameters and breeding values with higher accuracies. Based on the simulation results, the MF-SSGBLUP performed better than other classical genomic selection methods in joint breeding.

关 键 词：基因组选择 元共祖 基因组联合育种 一步法 模拟研究 

分 类 号：S823.91[农业科学—畜牧学]