Machine learning-supported cybergenetic modeling, optimization and control for synthetic microbial communities

Synthetic microbial communities are promising production strategies that can circumvent, via division of labor, many challenges associated with monocultures in biotechnology. Here, we consider microbial communities as lumped metabolic pathways where their members catalyze different metabolic submodules. We outline a machine learning-supported cybergenetic strategy for manipulating the reaction rates of microbial consortia via dynamic regulation of key biomass population levels. To do so, we show a quasi-unstructured modeling approach for synthetic microbial communities with external regulation of intracellular growth regulatory components. Then, we formulate an optimal control problem to find the optimal initial conditions and dynamic input trajectories. We use model predictive control to address system uncertainty, which can be coupled to an observer based on moving horizon estimation. Using a two-member community with optogenetic control as a simulation example, we found the optimal initial biomass concentrations and light intensity trajectories to maximize naringenin production.