| Three basic statistical methods: mixture model method (MM), regression on genotypic probability (RGP) and least squares method (LS) in QTL mapping were evaluated by simulation study. These results suggest that each of the three methods has its own strength. When detecting additive QTL, LS had the highest power; when detecting dominance QTL, MM and RGP had the higher power and LS had the lowest power for all heritability levels. In the estimation of location for additive QTL, LS had the best accuracy (smallest MSE), followed by RGP, with MM having the lowest accuracy. In the estimating of the location for dominance QTL, MM had the best accuracy. When estimating additive effect, LS had the best accuracy followed by RGP. For estimating dominance effect, the three methods had similar performance but the accuracy of MM decreases when heritability is low.; Dominance effect is a cause of heterosis. Marker contrasts are used for the detection of dominance effects and for the estimation of QTL location and the dominance effect under the F-2 and the reciprocal backcross designs. With these marker contrasts the dominance effect not only can be estimated invariant to additive effects but also can ensure the independence of estimation from other linked QTL. Analytical formulae for recombination frequency and dominance effects were developed for three cases: single QTL, side interval and middle interval. The simulation study showed that these formulae are useful statistical tools. It also showed that with conditional markers included in the model the accuracy of estimating QTL location will be improved while the accuracy of estimation for dominance effects will decrease.; In most studies the interference effect was assumed nonexistence among the loci. A simulation study showed that genetic sampling could cause false interference although there was no interference in the original population. If interference effect was considered in least square analysis the accuracy of estimating QTL location could be improved as much as half. |
