Introduction - Traditionally, quantitative genetics considers that the phenotype of a plant (P) is determined by its genotype (G), the environmental conditions where it grows (E), and the GxE interaction (P = G + E+ GxE; Via & Lande 1985). This formalism guided plant breeding for decades, yielding our artificially selected crops. Soil microbiota, often considered as a part of E, have also a renowned importance in defining the plant phenotype. Yet, it remains unclear when we should care for the effect of the soil microbiota on plant phenotypes, relatively to G and E. Our objective was to assess this relative importance of the soil microbiota (M). We tested the inclusion of M as a distinct factor in this formalism, in order to quantify its relative importance next to G and E (P = G + E + M + GxE + GxM + MxE + GxExM; Oyserman et al., 2021). Such a formalism would help identifying cases where the soil microbiota could be a significant action lever for improving plant phenotype, or not.

M&M - In a full factorial design experiment, we grew three different genotypic accessions Arabidospis thaliana (G: Can-0, Col-0, Cvi-0) in three different autoclaved soils with contrasting textures (E: sandy, loamy, clayey), inoculated with the three different soil microbial communities originating from these soils (M: sandy, loamy, clayey microbiota). Using variance partition analysis, we quantified how G, E and M affected two important plant traits: growth, and tolerance to the necrotrophic fungi Botrytis cinerea, the causal agent of the grey mold disease. The three genotypic accessions were selected based on their different susceptibility over B. cinerea (Can-0: highly susceptible > Col-0: mid-susceptible > Cvi-0: more tolerant; Denby et al. 2004).

Results - We showed that M ranked as the third significant source of variance in plant growth (6%, p < 0.001), behind G (10%, p < 0.001) and E (57.5%, p < 0.001). The GxExM interaction (3.1% p = 0.03) showed that part of the plasticity in plant growth can be attributed to M. This overall effect of M on plant growth resulted in a difference of 29% between the most extreme microbiota modalities. We found that the GxExM interaction (14.7%, p < 0.001) and GxM (5.6%, p < 0.01) together ranked second in explaining the tolerance to fungal infection, behind G (29.3%, p < 0.001). In the most extreme cases, the soil microbiota would explain a reduction of foliar necrosis size up to 2-folds.

Conclusions - Our results clearly indicate that the soil microbiota is a non-neglectable source of variance in explaining plant phenotypes. This effect, often referred to as “microbiability” could represent a promising alternative to the traditional genetic selection for improving plant phenotype under the current context of global changes.