Digital technologies have tremendous potential for facilitating and accelerating the development and deployment of agroecological innovations: optimized varietal selection taking into account plant-microbiome interactions, biostimulation and management of plant immunity, biocontrol based on the use of bioinputs or management of resident beneficial organisms, plant nutrition.
Agro-ecological cropping systems experience a wide range of biotic interactions with complex microbial communities: beneficial to the plant, providing important nutritional and biodefense functions, or detrimental, notably microbial parasites and pathogens exploiting plant resources. The diversity and dynamics of these interactions depend on ecological conditions, including the phenotypes of the interacting species, on physiology, and on the abiotic environment. Deciphering the links between interspecific diversity, community structure, and biological functions is key to understanding, maintaining, diagnosing, and exploiting the community dynamics underlying the health or illness of a crop, and adapting agroecological systems to environmental stresses.
We will design new multi-omic data analysis tools and develop multi-scale spatio-temporal models of microbial communities in crop plants. This will require significant development of algorithmic, dimension reduction, and machine learning methods for data analysis; significant advances in ODE and PDE based system dynamics, hybrid numerical and discrete modeling of the complex biological systems; and significant advances in transversal AI and HPC techniques. Acquisition of novel data to support AI machine learning and numerical modeling is necessary to provide methodological guarantees and validation. We will capitalize on existing culture systems to anchor our work to pertinent challenges in agroecology, while acquiring novel data specific to this PEPR.
We will use large-scale, genome-resolved metagenomic analysis to address key questions about the functional properties of plant-microbiome biochemical reaction networks. We will implement community-scale metabolic network analyses to identify key species and metabolic functions that mediate plant-microbiome interactions. Advances in AI will be needed to cope with noisy data as well as to refine knowledge-based reasoning methods for network analysis. We will construct digital twins of reduced microbial communities and plant–microbiome systems, through spatio-temporal models fitted to experimental data. We will use culturomics to design and cultivate microbial communities as controlled and repeatable experimental models of natural communities, for each pathobiont or symbiont system of interest. Plant-microbiome feedbacks will be characterized through the evolution of microbial communities and plant response.The two-pronged systems biology strategy of this project addresses methodological challenges up- and downstream of specific applications in biocontrol. An expected outcome of this proposal is development of computer software and mathematical tools for a deeper understanding of the links between microbial community structure and crop response. This entails identifying key drivers of plant microbial communities that impact plant health and crop traits to design and assess ideotypes and species associations of agronomic interest. New modeling tools will foster knowledge of crop-microbiome diversity and interactions, with translational applications to be tested in the framework of big challenges on plant biocontrol. The data acquired for this project will constitute a unique reusable resource for French scientists: those developing methods, those developing applications, and those representing agroecological stakeholders.