Cloning quantitative trait loci by insertional mutagenesis.

This study explores the theoretical potential of "insertional mutagenesis" (i.e., mutagenesis as a result of integration of novel DNA sequences into the germ line), as a means of cloning quantitative trait loci (QTL). The approach presented is based on a direct search for mutagenic effects of a quantitative nature, and makes no assumptions as to the nature of the loci affecting quantitative trait value. Since there are a very large number of potential insertion sites in the genome but only a limited number of target sites that can affect any particular trait, large numbers of inserts will have to be generated and screened. The effects of allelic variants at any single QTL on phenotype value are expected to be small relative to sampling variation. Thus two of three stages of replicate testing will be required for each insert in order to bring overall Type I error down to negligible proportions and yet maintain good statistical power for inserts with true effects on the quantitative traits under consideration. The overall effort involved will depend on the spectrum of mutagenic effects produced by insertional mutagenesis. This spectrum is presently unknown, but using reasonable estimates, about 10,000 inserts would have to be tested, at reasonable replicate numbers (n ≧ 30) and Type I error (α=0.01) in the first testing stage, to provide a high likelihood of detecting at least one insert with a true effect on a given quantitative trait of interest. Total offspring numbers required per true quantitative mutagenic effect detected decrease strongly with increased number of traits scored and increased number of inserts per initial transformed parent. In fact, it would appear that successful implementation of experiments of this sort will require the introduction of multiple independent inserts in the original parent individuals, by means of repeated transformation, or use of transposable elements as inserts. When biologically feasible, selfing would appear to be the method of choice for insert replication, and in all cases the experiments must be carried out in inbred lines to reduce error variation due to genetic segregation, and avoid confounding mutational effects of the insert with effects due to linkage with nearby segregating QTL. The special qualifications of Arabidopsis thaliana for studies of this sort are emphasized, and problems raised by somaclonal variation are discussed.