Assessment of DNA pooling strategies for mapping of QTLs.

The synthesis of "DNA pools" from segregating populations is an efficient strategy for identifying DNA markers closely linked to genes or genomic regions of interest. To-date, DNA pooling based solely upon phenotypic information, or "bulked segregant analysis", has been employed only in the analysis of simply-inherited traits. We have assessed the utility of phenotype-based DNA pools for "tagging" (e.g., identifying DNA markers closely-linked to) quantitative trait loci (QTLs), segregating in the presence of other such loci, and expressing phenotypes which are influenced by the environment. Theoretical estimates suggest that QTL alleles with phenotypic effects of 0.75-1.0 standard deviations (SD), or larger, should be detectable in back-cross (BC), F2 and recombinant inbred (RI) or doubled haploid (DH) populations of manageable size (100-200 plants/lines). However, post hoc analysis of three data sets, used in QTL mapping of tomato and rice, indicate that the majority of QTLs identified had allele effects of less than 0.75 SD, and thus could not be easily tagged in DNA pools. Segregation distortion can have a large effect on the allelic composition of DNA pools, necessitating the use of more individuals in the pools to minimize false positive and false negative results. In general, we suggest that use of phenotype-based DNA pools might be successful in tagging QTLs of very large effect, but is unlikely to permit comprehensive identification of the majority of QTLs affecting a complex trait. DNA pools constructed from a priori information should, however, be useful in identifying new DNA markers for regions of the genome known to contain QTLs.