M. oryzae is a phytopathogenic fungus causing important rice production loss. One reason for such a negative impact is that rice is cultivated in monoculture on very large field surfaces. This genetic homogeneity makes a perfect playground for a fungal pathogen that would overcome the plant's resistant genes. To counter the negative effect of fungal infections, farmers use pesticides, but this strategy has several pitfalls: first, these molecules are a threat to human heath and second, fungal strains eventually become resistant. An alternative strategy to fight fungal infection in the fields is to introduce genetic diversity in the rice population, and create a "mosaic landscape" where no unique fungal strain could thrive. During this project, we assessed the possibility of using variety mixtures to control M. oryzae in the field without the use of pesticides. Among other things, we studied the local adaptation patterns of M. oryzae on rice. We showed that M. oryzae is adapted to rice subspecies, and that these subspecies inherited different subsets of resistance genes. More specifically, we showed that there is a nested distribution of R genes among rice species. Such results can help in the choice of appropriate rice varieties to compose variety mixtures.