Résumé
Abstract Mastitis is a multifactorial infection of the udder potentially caused by many pathogens of varying severity and prevalence. They are one of the most common diseases on dairy farms. Limited information exists about the interactions between pathogens in mastitis infections. The first aim of this study was to investigate the dynamics of pathogens and their associations in the udder microbiota. We then explored the impact of these dynamics on mastitis risk, in order to gain a better understanding of this disease. We finally looked for factors influencing these dynamics and the potential sources of the studied pathogens at farm level. To address these objectives, two four-month longitudinal studies were conducted on cows of six dairy farms in the Auvergne region of France. Milk and faeces were collected from 33 cows, along with environmental samples (bedding and milk filter). The 16S DNA in these samples was analysed, using a commercial qPCR kit (PathoProof™) to detect 15 mastitis-causing pathogens. The data were then processed using principal component analysis, the Ward clustering method and discrete-time Markov chain models. We classified pathogens according to their distribution in the different compartments. Clustering analyses of milk quarter samples revealed distinct profiles of pathogen distribution associated with different somatic cell counts and cow recovery dynamics. According to these dynamics, C. bovis , though considered a minor pathogen, and S. uberis were pivotal in the definition of milk pathogen profiles, with the possible involvement of non- aureus staphylococci.