Identification and estimation of a QTL model and its effects

Abstract

A joint segregation analysis of a genetic system and the effects of QTLs based on the six populations P1, F1, P2, B1, B2 and F2 is proposed in this paper. The major steps were as follows. Firstly, under the supposition that the segregating population was composed of component distributions controlled by a major gene(s) and modified by both polygenes and environments, four groups and 17 types of genetic models, including a one major-gene model, a two major-gene model, a polygene model, and a mixed one-major gene and polygene model, were set up. Secondly, the joint maximum-likelihood function was constructed from the six generations so as to estimate the parameters of component distributions through an EM algorithm. Thirdly, the best-fitting genetic model was chosen according to Akaike’s information criterion, a likelihood-ratio test, and tests for goodness of fit. Fourthly, the related genetic parameters, including gene effects, as well as the genetic variances of major genes and polygenes, were obtained from the estimates of component distributions. Finally, the individuals in segregating populations were classified into their major-gene genotypes according to their posterior probabilities. An example of the genetic analysis of plant height of a rice cross between Nanjing No. 6 and Guangcong was used to illustrate the above procedure. The method was especially appropriate to those crops with easy to obtain hybrid seeds.