Additive main effects and multiplicative interaction (AMMI) analysis of genotype-location interaction in variety trials repeated over years

Genotype-location (GL) interaction effects are of special interest for breeding programmes to identify adaptation targets, adaptive traits and test sites. These effects, generally having relatively low repeatability between years, should be studied on a multiyear basis in annual crops. Their assessment by additive main effects and multiplicative interaction (AMMI) analysis is currently defined for this situation. Two procedures based on cross validations are proposed for testing the CL-interaction principal component axes, exploiting the utilities of the computer programme MATMODEL. The use of Gollob's F test, F-GH2 test, F-R test and the heuristic criterion based on the signal-to-noise ratio is also envisaged. The consistency of results provided by the testing procedures was verified on four data sets of different cereal crops. Gollob's test tended to be the most liberal, while the F-GH2 test appeared somewhat more liberal than the F-R test. The signal-to-noise ratio gave results consistent with the F-R test considered at a P less than or equal to 0.01 level of significance. These criteria disagreed in two data sets with the conclusions provided by the two cross-validation procedures which, in turn, also disagreed in one data set. Preference could be given to different testing procedures depending on the number of test years, locations and genotypes.