The use of either genomics or metagenomics data combined with metabolic modelling provides putative mechanistic understanding on microbes’ interactions within large scale microbial communities. Genome-scale metabolic networks (GSMNs) can be inferred automatically from large collections of annotated reference genomes or metagenome assembled genomes using dedicated reconstruction tools and can be used to infer the behavior of the underlying microbial community. Such community-scale modelling can predict exchanges of metabolites between species as witnesses of the cooperative or competitive potential in the community. To do so, numerical approaches, mainly based on flux balance analysis, characterize bacterial community with a good accuracy by comparing individual and community growth rates. Other approaches determine competition and cooperation scores from the topological analysis of networks. A shared constraint between methods is the limitation to pairwise analysis that prevents complex interactions in the community. In this talk, I will introduce a reasoning-based approach for the characterization of cooperation and competition potentials in large-scale microbial communities that is not limited to pairwise analysis. This method relies on modelling the metabolic potential in each microbe using a Boolean abstraction of metabolic producibility previously used for single organisms and ecosystem characterization. I will present a set of metrics that can be used to predict the potential for cooperative interactions and competition between species. We applied them to existing microbial communities and created benchmarks aiming at comparing our scores in several ecosystems. Results show variations in competition and cooperation potential in distinct ecosystems, and comparison to existing pairwise metrics evidence correlations between predictions. We believe such approach to be relevant in the journey towards the better understanding and control of complex microbial ecosystems.