QTL混合遗传模型扩展至2对主基因＋多基因时的多世代联合分析

本文将Gai and Wang(1998)的P1、 F1、 P2、 B1、 B2和F2 6个世代 联合分离分析方 法从A、 B、 C、 D 4类共17种遗传模型扩展至E(2对主基因＋多基因)共5类24种遗传模型 。 成分分布参数估计的方法由EM算法改进为迭代条件EM算法(IECM)， 其收敛性和所获结果 优于EM算法。 水稻株高例中2对主基因+多基因的遗传模型(E-2)优于原所

A Joint Analysis of Multiple Generations for QTL Models Extended to Mixed Two Major Genes Plus Polygene

The two major genes plus polygene Model (Model E ) of QTL in a joint analysis of P 1, F 1, P 2, B 1, B 2 and F 2 was extended to based on Gai and Wang (1998). The major steps were to establish the genetic models and respective maximum likelihood functions, to estimate the distribution parameters through iterated expectation and conditional maximization (IECM) algorithm, to select the best genetic model through AIC value and tests for goodness of fit, to estimate the genetic parameters of the best fitted model through least squares method, and to classify the individuals into major gene genotypes in terms of Bayesian posterior probability. The example of the inheritance of rice plant height which was used by Gai and Wang (1998) was further analyzed to explain the procedures. It was shown through the example that the convergence of the parameter estimation in IECM algorithm was better than that in EM algorithm, and the two major genes plus polygene model ( E 2 ) was better than the previous one major gene plus polygene model (D ).