利用最大似然法进行水稻产量性状基因的分子作图

本研究根据对估计标记－数量性状基因座位（ＱＴＬ）之间重组率的两种分析方法（矩量法和最大似然法）、两种方差模型（ＱＴＬ基因型之间的方差同质和异质模型）的分析，揭示了ＬＯＤ值在标记－ＱＴＬ连锁检测上所得结果的相关性高于重组率估计值的相关性。采用最大似然法和异质方差模型，估计了水稻产量构成有关的ＱＴＬ与分布于１１对染色体上的５１个限制性片段长度多态性（ＲＦＬＰ）标记之间的重组率，并对似然比（以ＬＯＤ值表示）进行Ｘ￣２检验，发现７个存在显著连锁关系的标记－性状组合，其平均重组率为１０．０％。这些标记分布于第１、５、６、８和１１等５对染色体上，涉及７个ＲＦＬＰ标记和３个产量构成性状，即每穗颖花数（ＲＧ５７３、ＲＺ６１７、ＲＧ１０３）、单株穗数（ＲＧ６４Ｂ）和每穗实粒数（ＲＧ１０１、ＲＧ２４４、ＲＧ６５３）。

Molecular Mapping for Quantitative Trait Loci Controlling Yield Component Characters Using a Maximum Likelihood Method in Rice(Oryza sativa L.)

For different methods(moment solution and maximum likelihood estimation) and different models(homoscedastic and heteroscedastic models), more significant correlation was found between LOD scores than between the estimators of recombinant fraction. Maximum likelihood method under homoscedastic modes was then used to estimate recombinant fractions between quantitative trait loci controlling yield components and 51 restriction fragment length polymorphism markers on 11 chromosomes of rice(Oryza sativa L.). Using X2 test for likelihood ratios(expressed as LOD scores), significant linkage relationship was found for 7 marker-trait combinations with mean recombinant fraction of 10.0%. These combinations involved 5 chromosomes(1,5,6,8 and 11), 3 traits and 7 markers, i.e.,spikelet number per panicle(BG573, RZ617, and RG103), panicle number per plant(RG64B),and grain number per panicle(RG101,RG244 and RG653).