三交组合ADAA模型和种子模型的Monte Carlo模拟分析

采用Monte Carlo模拟验证了三交方式遗传模型及其统计方法的可靠性．结果表明：(1)采用MINQUE(1)法可以无偏地估计ADAA模型中的各遗传方差分量,采用MINQUE(0／1)法可以无偏地估计种子模型中的各遗传方差分量;在种子模型中,当直接加性与母体加性效应间的协方差、直接显性与母体显性效应间的协方差不存在或不重要时,可利用MINQUE(1)法对缩减的种子模型进行重新分析,并获得无偏的分析结果。(2)分析ADAA模型或种子模型需要亲本、单交F1和三交F13个世代,在此基础上增加新的世代或扩大样本容量更有利于提高方差分量和遗传效应的估算(预测)效益。(3)ADAA模型或种子模型中各遗传效应分量可采用AUP法进行无偏地预测。

Monte Carlo simulation for analyzing ADAA models and seed models based on the triallel mating design

Monte Carlo simulations were performed to prove the efficacy of the genetic models and its statistic methods for three-way cross. The results indicated that:1) unbiased estimates of genetic variance components could be obtained by methods of MINQUE(1) for ADAA models and MINQUE(0/1) for seed models; if covariance between direct additive and maternal additive effects, and that between direct dominance and maternal dominance effects do not exist or are negligible, MINQUE(1) method can be employed for re-analyzing the reduced models to obtain unbiased results. 2) Three generations including parents, F_1 of single crosses and F_1 of three-way crosses were necessary for the analysis of ADAA models or seed models. The precision of the estimates and predictors improved with increase of number of generations and/or increase of the sample size. 3) Genetic effects could be predicted without bias by adjusted unbiased prediction (AUP) method for both ADAA models and seed models.