In this thesis, I investigated the relationship between host and pathogen in multi-host and multi-pathogen systems at the interface between wildlife and domestic species. The term "epidemiological interaction" was central to my thesis, and was defined as "any ecological interaction between two host populations resulting in the transmission of one or more pathogen". Epidemiological interactions are related to the processes of transmission between hosts and I investigated how these epidemiological interactions between different host populations could be investigated in a given ecosystem. I developed two research frameworks to estimate these epidemiological interactions: 1) an a priori approach based on the host data and assuming that the mobility of hosts and the resulting contacts between host populations would be crucial factors influencing the epidemiological interactions; 2) an a posteriori approach based on the pathogen data, assuming that epidemiological pathways previously used by some pathogen species can be used in the future by other pathogens. The animalpathogen model used to test the first approach was the bird-avian influenza viruses' model. Longitudinal counting and sampling protocols of domestic and wild birds over two years were used to analyse community composition and abundance of hosts to compare with the prevalence of avian influenza viruses. I could, for the first time, show a persistence of low pathogenic avian influenza strains in an African ecosystem, and investigate the relationships with both the potential maintenance hosts (Afro-tropical ducks and resident species) and hosts that introduced the virus into the system from Europe or Asia (paleartic migrants). With the estimation of epidemiological interaction using host community data, I estimated the contact rate between wild and domestic avian compartments (intensive poultry, backyard and farmed ostrich compartments) and assigned a risk to this interaction based on dynamic and non dynamic fa