Évaluation et prédiction des activités PGP par approches statistiques

Abstract

PGPR activities constitute a set of functional mechanisms involved in plant nutrition and the biocontrol of phytopathogens. In this work, 460 bacterial isolates obtained from the rhizosphere of 34 soil samples collected in the Mostaganem province were evaluated for 11 PGPR activities. The objective was to analyze the interdependencies between these functional traits and to model their predictability based on phenotypic profiles. The studied activities included phosphate and zinc solubilization, nitrogen fixation, production of IAA, siderophores, and exopolysaccharides (EPS), as well as the production of five hydrolytic enzymes (amylase, lipase, protease, cellulase, and pectinase). Correlation analyses (Pearson, Spearman, dCor) and PCA revealed overall weak links between most mechanisms, with the exception of a marked antagonism between EPS production and hydrolytic enzymatic activities, reflecting a metabolic trade-off in the allocation of carbon resources.. In contrast, IAA production remained consistently unpredictable (negative R² for all three approaches), suggesting an autonomous genetic regulation, disconnected from the other measured PGP traits.