Community diversity determines the evolution of synthetic bacterial communities under artificial selection

Artificial selection can be conducted at the community level in the laboratory through a differential propagation of the communities according to their level of expression of a targeted function (i.e. community phenotype). Working with communities instead of individuals as selection units brings in additional sources of variation in the considered phenotype that can arise through changes in community structure and influence the outcome of the artificial selection. These sources of variation could even be increased by manipulating species diversity. In this study, we wanted to assess the effect of manipulating initial community richness on artificial selection efficiency, defined as the change in the targeted function over time as compared to a control treatment without artificial selection. We applied artificial selection for a high productivity on synthetic bacterial communities varying for their initial richness level (from one to 16 strains). Our results showed that, overall, the communities that were artificially selected were 16% more productive than the control communities. Community richness positively influenced community productivity and metabolic capacities and was a strong determinant of the dynamics of community evolution. Our results suggested that community richness could influence artificial selection efficiency but a convergence of the community composition might have limited the effect of diversity on artificial selection efficiency. We propose that applying artificial selection on communities varying for their diversity could allow to find communities differing for their level of expression of a function but also for their responsiveness to artificial selection, provided that their initial composition is different enough.