Study On Methods For Mapping Epistatic QTLs Underlying Endosperm Traits Based On E-bays

Cereal crops are essential ingredients to human beings, especially to Asian people. As an important component of seeds in the cereal crop, endosperm is a rich resource, including protein, amino acid and starch. Therefore, it is necessary to conduct investigation on the endosperm traits in order to enhance the production of cereal crops and improve the quality. Unlike ordinary diploid plants, the genetics of endosperm traits is triploid formed when two polar nuclei combine with one sperm nuclei, which is more complicated. There are three genetic effects for endosperm traits:additive effect, first dominant effect and second dominant effect. Nowadays, different analytical approaches have been proposed by researchers for QTL mapping. However, QTL mapping and epistasis analysis for endosperm trait was still few.In this study, two research methods are proposed for QTL mapping of endosperm trait based on triploid. In BC1.1, BC1.2 and F2 population, investigations are conducted on QTL mapping for endosperm trait using the multi-variance-component model. Epistasis studies are carried out on the QTL of endosperm trait applying the Empirical Bayesian statistical methods, according to different prior distributions (inverse chi-square prior and index priori). The methods are verified by computational simulation. The main results are given as follows:1. Supposed that the random effect variances (additive effect, first dominant effect and second dominant effect) are equal under the same environment, the analysis of main effect for endosperm trait is conducted with the assumption that the random effect variances proportional to the environmental variance. A multi-variance component model is proposed based on the analysis. Simulation study shows that the methods can precisely map endosperm QTLs and unbiasedly and efficiently estimate the three effects (additive effect, first dominance effect, second dominance effect) of endosperm QTLs.2. The analysis of epistatic QTL controlling endosperm trait is conducted using Empirical Bayesian (E-Bays) statistical methods, according to two different prior distributions. The first study of QTL mapping is conducted using the E-Bayes based on the inverse chi-square prior distributions, considering the coordinate descent and block descent. The second study of QTL mapping is carried out using the E-Bayes based on the index priori distributions. Simulation studies show that the methods can efficiently map endosperm QTLs and precisely estimate their effects, no matter that the inverse chi-square prior and the index priori distributions.