一种适用于单胎动物多品种选择的基因组关系矩阵

旨在提出一种新型基因组关系矩阵并验证其在多品种联合群体中的模拟应用效果。本研究利用QMsim软件模拟牛的表型数据和基因型数据；利用Gmatrix软件构建常规G阵；利用R语言构建新型G阵，新型G阵在常规G阵的基础上，将多品种联合群体的非哈代-温伯格平衡位点考虑在内；利用DMU软件使用“一步”法模型计算基因组估计育种值（estimated genomic breeding value，GEBV）；比较不同情况下使用两种G阵的GEBV预测准确性。结果表明，在不同遗传力及QTL数下，不对新型G阵使用A₂₂阵加权就能达到常规G阵使用A₂₂阵加权时的GEBV预测准确性。在系谱部分缺失时，新型G阵不加权较常规G阵加权时GEBV预测准确性高。证明，在系谱有部分缺失时，新型G阵对多品种GEBV的预测有一定优势。

A Novel Genomic Relationship Matrix Applied to Single Birth Animals for Across-breeds Genomic Selection

This study aimed to propose a new genomic relationship matrix and validate its simulated application efficacy for an across-breeds population. The bovine phenotypic and genotypic data were simulated by QMsim software; General G matrix was constructed by Gmatrix software; The new G matrix was constructed by R language. Compared with the general G matrix, the new G matrix took into account the Hardy-Weinberg disequilibrium loci in combined population; The ssGBLUP method in DMU software was used to calculate the estimated genomic breeding values; The GEBV accuracies predicted by two G matrixes were compared for different situations. The results indicated that under different heritabilities and QTL numbers, the GEBV accuracy of the general G matrix using A₂₂ matrix weighting could be achieved without using the A₂₂ matrix weighting for the new G matrix. When the partial pedigree information missed, the GEBV accuracy of new G matrix without weighting outperformed general G matrix with weighting. In conclusion, new G matrix could perform better on GEBV prediction for across-breeds population when the partial pedigree information missed.