Separating environmental effects from those of biotic interactions on species distributions has always been a central objective of ecology. Despite years of effort in analysing patterns of species co-occurrences and communities and the developments of sophisticated tools, we are still unable to address this major objective. A key reason is that the wealth of ecological knowledge is not sufficiently harnessed in current statistical models, notably the knowledge on biotic interactions. Here, we develop ELGRIN, the first model that simultaneously combines knowledge on species interactions (i.e., the metanetwork), environmental data and species occurrences to tease apart the relative effects of abiotic factors and overall biotic interactions on species distributions. Instead of focusing on single effects of pairwise interactions, which have little sense in complex communities, ELGRIN contrasts the overall effect of biotic interactions to that of the environment. Using various simulated and empirical data, we demonstrate the suitability of ELGRIN to address the objectives for various types of interactions like mutualism, competition and trophic interactions. Data on ecological networks are everyday increasing and we believe the time is ripe to mobilize these data to better understand biodiversity patterns. ELGRIN provides this opportunity to unravel how biotic interactions actually influence species distributions.