考力代羊生长性状的协方差函数估计  

Estimation of Covariance Function for Growth of Corriedale Sheep

作　　者：李秀丽[1,2] 刘宝凤[2] 乔利英[2] 杨菊凤 张善芝 徐相亭 刘文忠[2] 

机构地区：[1]山东畜牧兽医职业学院,潍坊261061 [2]山西农业大学动物科技学院,太谷030801

出　　处：《中国畜牧兽医》2015年第1期167-171,共5页China Animal Husbandry & Veterinary Medicine

基　　金：山西省科技攻关项目(011029);山东省现代农业产业技术体系羊产业创新团队(SDAIT-09-011-05)

摘　　要：试验旨在模拟考力代羊生长过程中的遗传和表型变异,为其选种提供一定的依据。用1 269只考力代羊的3 332个生长发育记录,配合自变量为年龄的勒让德多项式的随机回归模型（RRM）,其中个体加性遗传效应配合5阶,母体永久环境效应配合2~5阶,对数约束最大似然值和信息标准用来检测最佳配合阶数。用平均信息约束最大似然（AIREML）法来估计体重的遗传、环境和表型协方差函数,结果表明,母体永久环境效应配合4阶时,协方差图更符合考力代羊的生长变化情况,涉及的参数也较少,且羔羊在6月龄前后表现出较大的遗传和表型变异。这些结果提示随机回归模型在有限的数据下能够充分描述考力代羊体重在连续时间轴上的增长变化情况。The purpose of this study was to model genetic and phenotypic variation of Corriedale sheep and provide reliable basis for selecting breed.Data set of 3 332 body weight records from1 269 Corriedale sheep was used to estimate genetic,environmental,and phenotypic covariance function by random regression model（RRM）,the independent variables were Legendre polynomial of age at recording.Orders of fit from 2to 5for maternal environmental effects were considered,and full order of fit for additive genetive effects.The best order of fit was tested by log likelihoods,Akaike and Bayesian-Schwarz’s information criteria.The average information restricted maximum likelihood method（AIREML）was used to estimate covariances between the regression coefficients.The results showed that when 4-order maternal environmental effects fitted,covariance figures were more in line with the change of the growth of animal,the less number of parameters involved.The greater genetic and phenotypic variation were showed before and after6 months.These results suggested the Corriedale sheep growth was adequatedly described using RRM despite the limited data.

关 键 词：考力代羊 经度数据 协方差函数 随机回归模型 平均信息约束似然法 

分 类 号：S826[农业科学—畜牧学]