Advantages of Multivariate Analysis of Profiles for Studies with Temporal Variation of Treatment Effects in Plants

Abstract

Repeated measures are widely used in a variety of aspects related to plant research, such as the assessment of genotypes subjected to stressing agents. Analysis of variance is routinely employed in split-plot experimental designs with repeated measurements on the subplots. However, this is not the best approach, as measurements cannot be randomly taken and so the H-F condition is rarely met. In this study, procedures of multivariate analysis of profiles (MAP) are presented as statistically advantageous options for solving similar analytical problems. We used data from trials with 10 clonal genotypes of cacao (Theobroma cacao) in which their seedlings’ resistance against ceratocystis wilt was assessed on 11 repeated measures that did not comply with the H-F condition. Multivariate analyses based on the disease-response profiles of all clones allowed us to understand the genotype-specific patterns of disease growth, identify when profiles reached the plateau, verify the levels of damage at this point, and assess coincidence and parallelism of the genotypic profiles. On the basis of various pertinent contrasts, MAP also provided robust criteria for genotypic classification into groups with different resistance phenotypes based on disease levels at plateau points and similarity levels of responses within and between groups. The possibility of addressing other biological questions through similar methods and contrasts is discussed.