运用混合线性模型定位复杂数量性状基因的方法

Mixed model approaches of mapping genes for complex quantitative traits

运用混合线性模型分析原理，提出了复杂数量性状基因定位的方法，可以分析ＱＴＬ复杂的遗传效应及ＱＴＬ×环境互作效应．采用混合线性模型随机效应的无偏预测方法，可以预测基因型值和基因型×环境互作效应值，再运用区间作图法或复合区间作图法间接分析ＱＴＬ的加性、显性遗传主效应及其与环境的互作效应，还能定位在特定发育阶段表达的ＱＴＬ．基于混合模型的复合区间作图法（ＭＣＩＭ法）可以分析多环境的遗传实验资料，直接分析包括上位性效应的遗传主效应及其与环境的互作效应．Ｍａｒｋｏｖ链蒙特卡罗（ＭＣＭＣ）分析方法可用于推断ＱＴＬ的统计特征．

New QTL mapping methods based on mixed linear model approaches were proposed for analyzing complex genetic effects and their interaction with environments. Unbiased prediction can be applied for predicting genotype effects and genotype × environment interaction effects, which can then be further used for mapping QTL or developmental QTL with genetic main effects and GE interaction effects by interval mapping or composite interval mapping approaches. Mixed model based composite interval mapping approaches are capable of handling genetic data derived from multiple environments and directly analyzing genetic main effects (including epistasis) and GE interaction effects. Markov chain Monte Carlo (MCMC) methods can be applied to make inference for the statistical properties of QTL.