Rice Genetics IV - IRRI books - International Rice Research Institute

Simultaneous QTL introgression and identificationOne of the most important objectives of QTL mapping is to apply MAS for geneticimprovement of quantitative traits. Although marker-aided backcrossing has beensuccessfully used to transfer specific target genes into an otherwise desirable genotype(Huang et al 1997, Sanchez et al 2000), few efforts on MAS for QTLs have beenreported (Tanksley and Hewitt 1988, Shen et al 2001). Theoretically, the accuracy ofthe genetic parameters (effect and position) of a QTL is highly correlated with themagnitude of its effect (Wang et al 1999). Thus, it is expected that MAS for QTLswith very large main effects should be effective even though it is advisable to usemarkers that closely flank the target region. However, marker-aided introgression ofa target QTL affecting a low-heritability trait requires the transfer of a large donorsegment of ~10 cM flanked by two informative markers in the target QTL regiongiven a relatively precise estimation of the QTL position (Visscher et al 1996).As discussed in the above sections, epistasis and QE interactions are propertiesassociated with most QTLs, which, unfortunately, have been largely ignored in mostQTL mapping studies. Also, because most parents in the previous mapping experimentsare not the parental lines used in common breeding programs, accurate MASfor QTLs cannot be practiced without information on QTL epistasis and QE interactions,both of which are difficult and very expensive to obtain. Thus, informationfrom previous mapping studies may not be used directly in designing MAS experimentsfor QTLs in breeding programs. Other than epistasis and QE interactions, manyimportant questions regarding QTLs remain unknown. For instance, for most traits,the number of polymorphic QTLs and the number of functional alleles at each ofthese loci in the gene pool are generally unknown. This is because almost all previousQTL mapping studies were conducted using biparental materials. Cross-populationcomparison can only provide limited information. In other words, the following questionsremain to be answered before any MAS experiment for QTL transfer can beused: (1) Do, and to what extent, the mapping studies provide reliable informationabout the parameters (locations, effects, and linked markers) of QTLs for target traitsto be used in breeding populations? (2) How many loci and alleles at each locus arethere and how great are their individual and combined effects for each of the targettraits? (3) Will the QTLs have the same phenotypic effects in the new genetic backgrounds(recipients) as the effects that were detected in the original mapping populationsor parents (presence of epistasis)? (4) Are detected QTLs associated with otherundesirable traits (genetic drag) and to what extent (genetic basis of genetic drag)?A new strategy of molecular breeding has been proposed and is being used atIRRI for simultaneous QTL identification and introgression (Li et al 1999c). Briefly,this strategy is defined as a comprehensive backcross breeding program involvingwell-designed and organized breeding activities of international collaboration, andintegration and use of DNA markers with phenotypic and genotypic (progeny testing)selection. The technical components of the strategy involve (1) inclusion of maximumgenetic diversity in the parental lines by selecting two core sets of the bestgermplasm from the primary gene pool; (2) molecular and phenotypic characteriza-QTL mapping in rice: . . . 167