基于显性模型的基因组选择中贝叶斯方法研究

发布时间：2018-04-13 08:19

本文选题：全基因组选择 + 贝叶斯模型　；参考：《畜牧兽医学报》2017年01期

【摘要】：本研究旨在探索显性效应对基因组育种值估计准确性的影响。基于贝叶斯A模型,根据加性效应和显性效应相关性,提出两种子模型:1)加性效应和显性效应相互独立的BayesAD1模型;2)显性系数(Dominant coefficients)与加性效应的绝对值相互独立,并且显性系数服从正态分布的BayesAD2模型。通过模拟数据比较加性效应模型BayesAD0和两种显性效应模型下基因组估计育种值(GEBV)的准确性,并且研究不同数量性状基因座(Quantitative trait locus,QTL)数、全同胞家系内个体数和加性方差与显性方差的比重对GEBV准确性的影响。结果表明,显性模型可以减缓随着世代变化GEBV的准确性降低的趋势。另外,显性方差的比重越大,对GEBV的准确性影响越大。当加性方差与显性方差之比达到0.25时,BayesAD2有较大优势,分别比BayesAD1和BayesAD0准确性高20.3%和28.4%。全同胞数越多,GEBV的准确性越高,QTL数目增多,GEBV估计的准确性随之下降。结果显示,对显性效应占较大比重即低遗传力的性状进行育种值估计时,考虑显性效应可以提高育种值估计准确性。
[Abstract]:The purpose of this study is to explore the dominant effects influence the accuracy of the estimation results of genomic breeding value. Based on Bayesian A model, according to the additive effect and dominant effect correlation, put forward two seed model: 1) BayesAD1 model of additive effect and dominant effect independent of each other; 2) (Dominant coefficients) dominant coefficient absolute value and the additive effect of each other independent and dominant coefficient obeys the BayesAD2 model of normal distribution. Through the simulation data between additive effect BayesAD0 model and two kinds of dominant effects model of genomic estimated breeding value (GEBV) and the accuracy of different quantitative trait loci (Quantitative, trait locus, QTL), effect of full sib families within the individual number and the additive variance and dominant variance proportion on the accuracy of GEBV. The results show that the dominant model with reduced accuracy can slow the generation changes the trend of the GEBV. In addition, the dominant party The greater the proportion of poor, affect the accuracy of the greater the GEBV. When the additive variance and dominant variance ratio reached 0.25, BayesAD2 has a larger advantage than BayesAD1 and BayesAD0 high accuracy 20.3% and 28.4%. full sib number, the accuracy of GEBV is higher, the number of QTL increased, the accuracy of the GEBV estimation is decreased. The results showed that the breeding value to estimate the dominant effect of a larger proportion of low heritability traits, considering the dominant effect can improve the estimation accuracy of breeding value.

【作者单位】：中国农业科学院北京畜牧兽医研究所;
【基金】：基金项目现代农业(肉牛)产业技术体系岗位科学家(CARS-38) 中国农业科学院科技创新工程-牛遗传育种(ASTIP-IAS03)
【分类号】：S813.1

【相似文献】