Tropical forests are hyper-diverse communities: hundreds of tree species can coexist within a single hectare of forest. Several mechanisms, such as niche partitioning or the Janzen-Connell effect, have been proposed to explain how so many species can stably coexist while competing for a limited number of resources. Among these mechanisms, the role of intraspecific variability (IV), which is large in tree communities, has only been recently considered. IV can result from an endogenous genetic variability that could make species less different, hindering their stable coexistence. A different view is that observed IV is primarily the result of an exogenous environmental variability operating at fine scale, which, when well accounted for, can actually reveal differences among species on unobserved dimensions, promoting species coexistence. The studies that have so far explored the effect of IV on species coexistence have been limited in scope, focused on species-poor systems, used disparate approaches, and reached contrasting results: IV could either promote, hinder or have no effect on species coexistence. Our project proposes to provide a clear synthesis and framework to disentangle and test the different underlying hypotheses, fostering a fundamental understanding of the effect of IV on species coexistence in hyperdiverse communities. To do so, we will combine different approaches including a literature review, analyses of empirical datasets, and both theoretical and data-based models. In particular, we will mobilize different tropical forest simulators that are able to simulate the high level of species richness these ecosystems shelter using different data-based parameterization to test the different hypotheses. In doing so, we will go beyond the previous theoretical approaches restricted to simplified systems, establish a direct link between various field data, and shed light on the urgent research issue of understanding and forecasting biodiversity and ecosystem functions under changing environment.