Batch-to-Batch Optimization with Model Adaptation Leveraging Gaussian Processes: The Case of Optogenetically Assisted Microbial Consortia

Microbial consortia are promising biotechnological production systems with the potential to divide complex metabolic pathways into smaller submodules, as well as make products and consume substrates that monocultures cannot. Maintaining optimal cell population levels and preventing mono culture formation challenge bioproduction by microbial consortia. Optogenetics allows for regulating the expression of key growth-regulatory genes through light to modulate cell population levels. Model-based dynamic optimization can determine optimal light trajectories and inoculum sizes that maximize product synthesis while satisfying system constraints, e.g., safety, economic, or technical aspects. Further-more, closed-loop dynamic optimization can address system uncertainty to a certain extent; however, its implementation is challenging due to limited online sensors. Alternatively, here we propose to perform open-loop optimization with batch-to-batch model adaptation based on Gaussian processes for maximizing bioproduction by optogenetically assisted consortia. The proposed approach enables knowledge transfer from existing to new models, improving predictability and optimization performance in each batch while avoiding costly and time-consuming modeling experiments. Compared to closed-loop optimization, this strategy is easier to implement as it does not rely on online monitoring, contributing to the state of the art in optimizing bioproduction by microbial consortia. We outline the applicability of the approach using simulation experiments of an optogenetically assisted consortium for the biosynthesis of the flavonoid naringenin, considering both parameter and model structure uncertainty.