Cow mastitis, an inflammation of the mammary gland, is responsible for animal suffering, economic loss and a high consumption of antibiotics in dairy farms. This multi-factorial disease is influenced by pathogen infection, udder microbiome, host factors and farm management practices. Staphylococcus aureus, one of the main mastitis-causing pathogens (with Streptococcus uberis and Escherichia coli) is particularly difficult to control once in the udder, promotes co-infections, and presents zoonotic risks. S. aureus chronic mastitis is the primary reason for early culling, as no other management strategy currently appears efficient. While S. aureus can be found in most environments, in some farms it becomes prominent while in others it appears naturally controlled. As any species, it is part of an ecosystem including competitors and natural enemies which could hamper its development either in the udder or in the farm environment. EcoSA aims to study its circulation, interactions and microbial ecosystems, to identify situations that facilitate its control at the farm level. More generally, investigating dairy farms microbial ecosystems is necessary to understand the relative importance of pathogen infections, biological interactions and an imbalance of the microbial communities (dysbiosis) on mastitis occurrences, and to be able to devise new management methods for this complex disease. Our proposal plans to describe the dynamics and circulation of the main mastitis-causing pathogens in cow’s udder and gut microbiota, farm beddings and milk filter; to characterize the microbial dynamics in these compartments to identify biological interactions and dysbiosis associated with mastitis and S. aureus prominence; to evaluate pathogens and communities response to perturbations due to farming interventions (antibiotic treatment, change in feed, cleaning); and to improve management of S. aureus and mastitis by proposing new ecosystem-based management procedures. In practice, we will sample milk, feces, bedding and the milk filter and record farming interventions in commercial farms with high or low S. aureus abundance during the 20 weeks period spent in housing (where mastitis incidence is the highest) for two years. We will use metagenomics, bacterial cultures, qPCR and somatic cell counts to detect commensal bacteria, pathogen abundances, bacteriophage abundances and the host immune response across this period. Data will be used to infer the parameters of a predictive dynamical model representing microbial growth and interactions, perturbations and immigration, which will support the investigation of pathogen dynamics and circulation, microbial communities states and state shifts associated with S. aureus prominence, and the respective roles of dysbiosis, de novo infections and farming interventions in mastitis. Based on dynamical modelling and dialog with dairy farmers, we will propose management methods for mastitis evaluated in silico and tailored to their socio-economic constraints, which could take the form of a surveillance system for mastitis pathogens and ecosystem shift, bacterial augmentation by oral route, topical application or the environment, and/or recommendations on farming interventions to protect or restore beneficial microbial communities. We expect the results of this project to provide breakthroughs in the understanding of mastitis, in microbial ecology and dynamical modelling; but also to have major health and socio-economic impacts by reducing the need for antibiotic usage in dairy farms, improving farm animal health and wellbeing and the farmer’s economic security. If successful, our innovative proposal could be a precursor to the development of new agro-ecological farming solutions to pathogen control, helping farms transition to sustainable farming practices in line with societal demand for more ethic breeding systems.