Mapping Method Of Quantitative Traits Loci Underlying Endosperm Traits In Cereals

Endosperm traits are the kind of important traits determining the grain quality in cereals. The endosperm, which is derived from two polar nuclei fusing with one sperm, is a triploid tissue whose genetic constitution is more complex than common diploid tissue. Although the development and the application of the quantitative genetic models for endosperm traits have made an advance in the inheritance research on the quality traits of cereals, mapping of quantitative trait loci (QTL) controlling the endosperm traits has not been given sufficient consideration until several years ago. Despite some statistical methods for interval mapping of QTL underlying triploid endosperm traits have been developed, two problems have still been in existence. Firstly, the ordinary least squares based method for endosperm traits has not paid attention to the variance heterogeneity and the mixture distribution of QTL genotypes within flanking marker genotypes; Secondly, the common used methods have not been available to estimate two dominance effects of endosperm traits.Based on the quantitative genetic models for triploid endosperm traits, two new maximum likelihood methods implemented via EM algorithm for mapping QTL underlying endosperm traits were proposed in this paper. The method I uses the DNA molecular marker genotypes of each plant in segregation population and the single endosperm observation of a few endosperms of each plant as data set to analyze QTL. The method II uses the DNA molecular marker genotypes of each plant in segregation population such as F2, backcross etc. and the plant average of several endosperms of each plant as data set for QTL mapping. In addition, the method II includes the exact and the approximate EM algorithms.Efficiency and feasibility of the methods have been verified through Monte Carlo simulation studies. The simulation results show as follows:(1) The above two methods provide accurate estimates of both the QTL effects and locations with very high statistical power. For example, in a F2 population with only 50 plants and 5 endosperm observations each plant, the statistical power of QTL detecting by method II is above 85%, whereas that of the method I is high to 92% for a small size QTL only with 10% of heritability.(2) A larger heritability and sample size will lead to better estimates for both two methods. In general, when sample size is larger than 100 plants and the number of endosperms per plant is above 10, the QTL positions and effects can be estimated accurately and precisely even for a QTL of 10% heritability.(3) For an appropriate sample size, the method I not only estimate the QTL locations precisiely but also is indeed able to dissect the two dominance effects of endosperm traits.(4 )While the exact EM algorithm of the method II can not significantly increase the statistical power of QTL detecting and the accuracy of QTL locating compared with the approximate EM algorithm, it apparently increase the precision of QTL locating. The exact EM algorithm has less the standard deviations of the estimated QTL locations than the approximate EM algorithm for all of the 48 treatments simulated in the paper. Moreover, the advantage of the exact EM algorithm in the precision of QTL position estimate over that of the approximate EM algorithm will be more apparent as QTL heritability and sample size increase.