Statistical methods for detecting major genes of quantitative traits using phenotypic data of a diallel mating

The objectives of this thesis are to develop and evaluate the statistical methods to detect the major genes controlling quantitative traits in tree breeding population using simulated data and phenotypic data derived from a half-diallel mating design.; Three simple statistical tests, the Bartlett test, the Fain test, and the log-ANOVA test, were evaluated for detection of major gene segregation in a half-diallel mating design. Powers of detection and robustness to the presence of a half-sib relationship, a skewed distribution, and unbalanced data were investigated by using a series of simulated data sets. The most powerful tests were the Bartlett test and log-ANOVA test, but the Bartlett test was sensitive to the presence of a half-sib relationship and a skewed distribution. The Fain test had very limited power to detect the segregation of major genes although it was robust to the presence of a half-sib relationship and a skewed distribution. Overall, the log-ANOVA test may be the best simple test for detecting the major gene segregation for a half-diallel mating dataset. The major gene effects and the possible genotypes of parents and progenies cannot be estimated by the simple tests.; A Bayesian based segregation analysis method was developed to detect a major gene having large effects on the phenotype of a quantitative trait. A mixed inheritance model with a major gene plus polygenes was used to analyze the progeny population derived from a half-diallel mating design. Three simple tests and the parent block Gibbs sampling segregation analysis were also applied to eight progeny data sets of second generation of loblolly pine (Pinus Taeda L.) from the North Carolina State University - Industry Tree Improvement Program.; The log-ANOVA test was the most suitable simple test for screening populations for possible major gene segregation with the diallel mating progeny data in this study. The Bayesian based segregation analysis may be more powerful to estimate the genetic parameters of the major gene and polygenes, and the possible genotypes of the major gene of parents and progenies. In combination with molecular confirmation, these statistical methods are invaluable for detecting major genes of quantitative traits. (Abstract shortened by UMI.)...