不平衡巢式设计遗传模型分析

Analyses of Genetic Models for Unbalanced Nested Design

虽然巢式试验设计在动植物育种中得到广泛应用,但是在不平衡数据条件下有关遗传参数估计和假设检验统计量的计算还存在着很多问题。文中针对不平衡数据条件下的巢式试验设计,使用固定效应模型和随机效应模型估计相关的遗传参数并进行相应的统计假设检验。对于固定效应模型,使用约束线性模型方法推导出亲本配合力估计及亲本间配合力假设检验统计量的计算公式。对于随机效应模型,采用混合线性模型中的方差分析法,推导出方差分量估计的计算公式,并给出方差分量估计标准误以及方差分量假设检验统计量的计算方法,进而给出遗传力计算公式及其标准误的近似计算方法。最后,用VC++编写多种形式巢式设计遗传模型的各种遗传参数估计和假设检验统计量计算的Windows应用软件,供林木遗传育种工作者使用。

Nested design has been widely used in animal and plant breeding, but there still exist some problems in calculating the estimates of genetic parameters and the statistics of some hypothesis tests with unbalanced data. In this paper, both fixed and random linear models were applied to estimate the genetic parameters and to perform tests of corresponding hypotheses with unbalanced data from the nested experiment design. For the fixed-effect model, we applied the methods of restricted linear model to estimate the general combining abilities of parents and to calculate the statistics for testing whether or not the general combining abilities of any two parents are different. For the random effect model, we used the theory of variance component model to derive the formulae for computing the variance components, their standard errors, and the statistics for hypothesis tests. Furthermore, formulae for calculating the heritabilities and estimating their standard errors were also given. Based on these results, a Windows-based software was developed to implement the calculations of all the genetic parameter estimates and the statistics for various hypotheses, and it would be a useful tool for forest breeders.