AFLP增效和减效位点预测杂交水稻产量性状模型构建

Foundation of Predictive Model of Hybrid Yield Traits in Rice (Oryza sativa L.) with Effect-increasing and Effect-decreasing Loci of AFLP

选用13个水稻不育系和19恢复系按NCⅡ设计配制2套半双列杂交组合,分别用每套组合F1产量性状对该套亲本AFLP多态性位点(总位点)进行筛选,获得阳性、增效和减效三类位点。分别用基于总位点、阳性、增效和减效4类位点的亲本遗传距离构建AFLP标记预测模型,对另一套组合产量性状进行预测。结果表明:(1)总位点中只有少数为阳性位点。与总位点相似,阳性位点用于产量性状预测时,预测值与实际值的相关系数(以下简称为预测系数)太低;(2)而单独利用增效或减效位点,以及同时利用该2类位点(构建二元模型),3类预测效果均大大提高,其中增效位点对结实率,增效或减效位点,及同时利用增效和减效位点对单株产量,4个预测系数在0.6以上,可以用于指导水稻育种;(3)利用2套组合均筛选到的共同增效和减效位点的预测也达到较高水平,其中4个预测系数在0.5以上,因其所使用的预测位点数大大减少,应用更为有利。

The positive, effect-increasing and effect-decreasing loci （PL, IL, DL）, which were screened from the parental AFLP polymorphic loci （TL） by the yield traits of two sets of half-dialell mating crosses designed used 13 cytoplasmic male sterile lines and 19 restorers of rice, respectively, together with TL, were utilized to estimate parental genetic distances （GDs）, and the models were founded to predict the traits of hybrids derived from different parents. The results were as follows： 1） PL occupied a few number of the TL, and its correlations between the predictive and authentic values of yield traits （termed as briefly predictive coefficient as following） were too low, as same as TL. 2） But the predictions by IL, DL or IL/DL （founding duality models） were increased significantly, e.g. the predictive coefficients were all above 0.6 such as IL for grain setting percentage, IL, DL or IL/DL for grain weight per plant, therefore, these models could be used as prophecy of rice breeding. 3） In addition, the predictions based on IL, DL or IL/DL from intersection of two sets of crosses were higher enough, four of which above 0.5, and was more suitable in molecular breeding due to less predictive loci required.