Résumé
Growing environmental constraints and the need for more sustainable food systems promote crop diversification and the harnessing of ecological processes to enhance resilience and reduce external inputs. Among the promising strategies, intercropping—growing multiple species together-optimizes resource use and fosters beneficial plant-plant interactions (Cougnon et al., 2022). However, designing productive and stable species mixtures requires both careful management of within-species variability and a better understanding of the genetic mechanisms underlying such interactions. Plant-plant interactions are major drivers of individual performance and community outcomes. In crop mixtures, they include both intra- and interspecific processes that shape competition, facilitation, and resource allocation. While their ecological consequences have been widely studied (Becker et al., 2023), their genetic basis—and especially their potential to evolve—remains underexplored. We address this gap using a model intercrop composed of Triticum durum (wheat), an annual income species expected to enrich grain protein content, and Medicago sativa (alfalfa), a perennial service species contributing to weed control and nitrogen input. This cereal-legume combination offers a compelling system to investigate interactions between species with complementary functions and life histories. To dissect these interactions, we rely on an individual-centered experimental design, where focal wheat plants are surrounded by variableneighbors from both species. This layout captures the intra- and interspecific genetic context in which each individual develops. By manipulating the genetic composition of neighbors, we analyze how their identity and diversity shape the phenotype of focal plants. We apply a quantitative genetic model estimating direct genetic effects (DGEs)—due to an individual’s own genotype—and indirect genetic effects (IGEs)—due to the genotypes of surrounding individuals (Bijma et al., 2007). Our model extends these concepts across species boundaries, capturing how a wheat genotype affects neighboring wheat and alfalfa plants, and vice versa. This is especially relevant in cereal- legume systems, where traits like nitrogen fixation or canopy architecture drive strong facilitative interactions. By partitioning phenotypic variance into intra- and interspecific genetic components and estimating covariances among them, we gain insights into how selection may act not only on individual performance but also on the ability to influence or respond to neighbors. Across functional traits (e.g., tiller number, plant height, leaf nitrogen), agronomic traits (e.g., biomass, yield components), and spectral data (NIRS of leaves and grains), we show that phenotypes are shaped by both self-genotype and the genotype of multiple neighbors. We demonstrate the existence of heritable interactions at both intra- and interspecific levels. For wheat, we estimate pseudo design heritabilities for DGEs and IGEs of 0.57 and 0.29 (plant height), and 0.52 and 0.26 (leaf nitrogen), confirming the capacity of our design to detect both. Simulations suggest these effects can evolve over time, shaping trait trajectories and breeding outcomes. By formalizing multispecies genetic interactions in an individual-centered framework, our work bridges quantitative genetics, evolutionary biology, and agroecology. It offers a novel lens to study cooperation and competition in crop mixtures and a practical toolbox for their design. This approach opens new perspectives: in breeding, by accounting for neighbor effects in both pure and mixed stands; in evolutionary biology, by exploring how interspecific interactions drive trait coevolution; and in agroecology, by illuminating the mechanisms behind facilitation and complementarity.