While the primary role of metabolism is chemical conversions, can it also serve as an information processing device? To answer this question, we propose to encode various microbial metabolic models into Artificial Metabolic Networks (AMNs), which can be trained on experimental data or model simulations. Unlike “black box” artificial neural networks, our AMNs will be sparse and will reflect faithfully the structure and dynamics of metabolic networks. Our AMNs will be benchmarked on classical machine learning problems to assess what level of computational sophistication metabolism is able to handle. In the context of biotechnology, our AMNs will be applied to the design of experiments to (i) optimize the productivity of an added-value chemical (lycopene) E. coli producing strain defining nutrient compositions and gene deletions and (ii) classify infectious disease severity by engineering an E. coli biosensing strain detecting metabolic biomarkers in COVID-19 clinical samples.

The importance of heterogeneous shedding patterns in the context of infectious diseases is now well documented and recognized. The infected individuals that harbour and shed a given pathogen at higher concentrations than their congeners are often referred to as super-shedder, by opposite to low-shedder individuals. These super-shedders have a much higher transmission rate and thus constitute a key target for epidemiological investigation and management of diseases such as salmonellosis for which poultry constitute the major source of human contamination. However, the conditions that favour their super-shedding phenotype are poorly understood but are a prerequisite to control the reservoir of contamination within a population. As the emergence of the super- and low-shedder phenotypes are determined by the gut microbiota present before infection, and have been observed in various animal species reared in distinct environments and in strikingly diverse gut microbiota compositions, we hypothesized that while different bacterial taxa may lead to similar outcomes in terms of heterogeneous shedding, potential functional and taxonomic commonalities in the intestine should lead to these phenotypes. Thus, based on preliminary data obtained in chicken, heterogeneous shedding appears to depend on a combination of (1) specific gut microbiota features (2) mucosal immune responses parameters (3) a complex metabolites-driven dialogue between host, pathogen and microbiota and (4) several stochastic effects, including the pace and success of the gut colonization by environmental micro-organisms. In contrast, we have shown that host genetics and modification of bacterial virulence do not play a major role. In this project, we will study the causes of the Salmonella Enteritidis heterogeneous shedding in chicken. Based on different conditions known to favour one of these phenotypes by modifying the gut microbiota composition, we will compare the specific gut microbiota features, the mucosal and systemic immune response parameters and the complex metabolites-driven dialog in the intestine. The MOSSAIC project, organized in six tasks, has three main objectives: To provide a deeper and more integrated understanding of the heterogeneous shedding phenomenon; To model the interactions of the partners of the biological “ménage à trois”: Salmonella-host immune response-gut microbiota, taking into account their metabolites; To confirm by using several in vitro and in vivo experiments some hypotheses suggested by the data analyses and mathematical models. The feasibility of our project is based on an original model of infection in isolator, which allow us to clearly identify the super- and low-shedder phenotypes by controlling animal cross contaminations. Moreover, the MOSSAIC project brings together 5 partners with complementary expertise required to carry out the work program: bacteriology, animal infection, avian immunology and metabolism, metabolomics, bioinformatics and modelling. In a long-term perspective, the knowledge gained during the project will serve as a basis for the development of bacterial communities, which could be fed to chicks to standardize their gut microbiota and increase their resistance to pathogens in poultry production sector.

The increasing consumption of fermented foods and beverages is among one of the more notable diet transitions observed in our societies. However, the strong demand for clean label food (rejection of chemical preservatives) and for sustainable agricultural practices (e.g. change towards organic production with low input of fertilizers and pesticides) leads to unforeseen modifications in the quality and safety of these food, especially those made from plant products (vegetables and fruits). Some raw materials are also undergoing biochemical modifications (e.g. in sugar content) due to climate change and increased water stress. These changes in raw materials lead to unexpected modifications in microbial ecology that can be detrimental to the fermentation processes and to the quality of fermented foods and beverages. In this context, the development of generic scientific approaches to help in understanding and anticipating the effects of multiple and complex changes in these productions is a real and urgent need. In addition, solutions to tackle these changes must be sustainable, like the exploitation of taxonomic and functional biodiversity of microorganisms. We will apply an approach of multi-omics-analysis and modelling of the food transformation ecosystems to two typical fermented foods: wine (liquid) and vegetables (solid), for which different challenges exist. These include low-alcohol wine production in a context of climate change, together with organic production and sulphite reduction, and, for various fermented vegetables, addressing health, safety and quality issues in a context of changing production scale (household versus semi-industrial), organic production and salt reduction. The goal of the project is to develop a knowledge-driven approach using synthetic ecology that aims, through the reconstitution of model foods, to predict the behaviour of microbial communities under different constraints. The objective of this strategy is to demonstrate that the production of meta-omics data (gene expressions at the ecosystem scale; global analyses of metabolite production) organized in microbial ecological networks by computational approaches is relevant to anticipate the impacts described above. Furthermore, the project will take advantage of the biodiversity of microbial strains available in partner’s collections for the construction of tailor-made microbial consortia based on the functional properties required to adapt to the expected ecological network changes. Finally, our hypothesis and the solutions obtained on model and simplified foods will be tested at the pilot scale of fermentations of real food to evaluate their validity and the organoleptic and/or nutritional relevance. In a context where fermented foods are at the heart of a societal (food production and healthy nutrition) and environmental (sustainability of production and processes) transition, the results of the METASIMFOOD project will have three major impacts. We will demonstrate the relevance and efficiency of the scientific discipline of synthetic ecology to predict, anticipate and control the behaviour of microbial communities involved in food fermentation processes. Our project will serve as a lever for the deployment of this strategy to other examples of food. It will provide the necessary knowledge base for the development of downstream research programs with industrial partners in the field of fermented foods. Finally, we will ensure a very strong communication towards the general public with an educational and popularization objective.

As the first available prebiotics for neonates, milk oligosaccharides regulate gut microbial composition and modulate host immune response, playing a crucial role in the holobiont assembly. By using two livestock models (pigs and rabbits) with different maturity levels at birth, HoloOLIGO aims to decipher causal links between milk oligosaccharidesstructures, the offspring microbiota and immune system. We will create a database using data mining of the literature to find, visualise and analyse milk oligosaccharidesstructure diversity patterns within and between mammalian species. We will produce the first MO data in rabbits and expand them in pigs. To understand structure importance of milk oligosaccharides, we will undertake in vitro functional analyses in both species on commensal bacterial strains and intestinal immune cells and further validate results in vivo. Finally, we will evaluate, via in silico analyses (pig) and in vivo (rabbit), the existing genetic variability and assess the genetic determinism of milk oligosaccharidescomposition.