Detection of QTL x environment interaction in maize by a least squares interval mapping method

In order to detect the linkage disequilibrium existing between alleles at a marker locus and alleles of a linked quantitative trait locus (QTL), a least squares interval mapping approach using multiple regression on marker data has been developed. It allows inclusion in the model of the parameters describing the experimental and environmental situation, so that the QTL x environment effects can be tested. The method can also be applied using any general statistical package to data for which the usual normal distribution assumption does not hold, and where the use of weighted approaches is therefore required. A method to cope with the frequent problem in biological experiments of missing data was also used. The analysis was performed on data concerning two components of maize pollen competitive ability, obtained from an experiment over 2 years. The method, in comparison with the traditional single marker approach, has been shown to be more powerful in detecting QTLs and more precise in determining their map position. The analysis has identified QTLs expressed across years, putative QTLs with major effects and QTLs accounting for genotype x environment interaction.