Abstract
The transmission disequilibrium test (TDT), originally developed for mapping disease genes, has recently been extended to the identification of quantitative trait loci (QTLs). A general TDT (TDTG) has been developed to test linkage and/or linkage disequilibrium between a QTL and a marker in nuclear families that may be composed of more than one heterozygous parent and multiple children. In the investigation of the statistical properties of the TDTg: 1) background polygenes were not considered, which is not realistic, for most quantitative traits are polygenic, and 2) all family members were randomly ascertained. Using a random sample in this case may have limited power, especially when the heritability is small. Usually only extreme values are of primary clinical interest and extensive records are accumulated. In this study, an analytical method was developed to compute the power of the TDTG under various degrees of contribution of background polygenes to phenotypic variation. It is shown that the power of the TDTG is increased with the effects of polygenes when more than one child is employed. How sampling of parents or children based on their extreme values would affect the power of TDTG was also investigated. Three situations were considered based on the selection of phenotypic values of parents or children: 1) two children, one is of extreme value, the other random; 2) two extremely discordant children; 3) one parent is of extreme value. The study shows that the above selective sampling schemes of family members can enhance the power for QTL identification. The increase in the statistical power is particularly dramatic when /?-7 at the QTL is small or intermediate. Therefore, clinical records of extreme individuals are powerful resources for QTL identification by the TDTG. The accuracy of the analytical methods was validated by computational simulations. The results of this study should be of theoretical significance in generalizing the investigation of the TDTG to biologically plausible situations. They should also be of practical value in how to perform the TDTG efficiently by taking advantage of extensive databases enriched with people of extreme phenotypic values.