Massey, C.; Nosker, M. E.; Gale, J.; Scott, S.; Walker, C. J.; Cluff, A. J.; Wilcox, S.; Morrison, A.; Morgan, S. J. G.; Beltz, J.; Schmidt, P.; Chaston, J.

The fruit fly Drosophila melanogaster is a model for understanding how hosts and their microbial partners interact as the host adapts to wild environments. These interactions are readily interrogated because of the low taxonomic and numeric complexity of the flies\' bacterial communities. Previous work has established that host genotype, the environment, diet, and interspecies microbial interactions can all influence host fitness and microbiota composition, but the specific processes and characters mediating these processes are incompletely understood. Here, we compared the variation in microbiota composition between wild-derived fly populations when flies could choose between the microorganisms in their diets and when flies were reared under environmental perturbation (different humidities). We also compared the colonization of the resident and transient microorganisms. We show that the ability to choose between microorganisms in the diet and the environmental condition of the flies can influence the relative abundance of the microbiota. There were also key differences in the abundances of the resident and transient microbiota. However, the microbiota only differed between populations when the flies were reared at humidities at or above 50% relative humidity. We also show that elevated humidity determined the penetrance of a gradient in host genetic selection on the microbiota that is associated with the latitude the flies were collected from. Finally, we show that the treatment-dependent variation in microbiota composition is associated with variation in host stress survival. Together, these findings emphasize that host genetic selection on the microbiota composition of a model animal host can be patterned with the source geography, and that such variation has the potential to influence their survival in the wild.