Factorial regression analysis of genotype by environment interaction in ryegrass populations, using both isozyme and climatic data as covariates

A sample of 58 natural ryegrass populations (Lolium perenne L.) from the French and Spanish oceanic coast was evaluated for three seasonal growth traits (i.e. spring, autumn and winter vigour traits) in 5 evaluation sites (three in France and two in Spain). This sample showed a high level of genotype (population) × environment (evaluation sites) interactions for the three agronomic traits. A factor regression analysis, using both isozyme frequencies of populations and climatic factors of evaluation sites as covariates, was carried out on a sub-sample of 30 populations in order to explain these interactions. This method succeeded in explaining most population × location interactions by the product of two covariates. For instance, for spring vigour trait, 72.8% of the interaction term could be explained by the use of two covariates: the PGI2-20 allelic frequency of populations and the minimum temperature of the coldest month of evaluation sites. This study shows the advantage of such a method for plant breeders who need to identify promising ryegrass populations for their breeding objectives. A number of genetic and evolutionary hypotheses are also discussed. This revised version was published online in July 2006 with corrections to the Cover Date.