Batch control of genetic alterations for optimal metabolic engineering

Metabolic engineering manipulations can be performed in an optimal manner to maximize desired cellular properties. In prior work [Gadkar, K. G., et al., 2004], a bilevel optimization framework was developed to demonstrate that temporal genetic manipulations yield optimal productivity. In this work, the bilevel optimization framework is coupled with control algorithms to determine the genetic manipulation strategies in practical bioprocess situations. Ethanol production in an anaerobic batch fermentation of Escherichia coli in two case studies are considered. In the first, the bilevel optimization framework is augmented to incorporate a penalty for longer time of operation. The framework successfully optimizes the batch time along with the genetic manipulations for maximizing the desired objectives. In the second, the bilevel optimization framework is coupled to a parameter estimation algorithm to compensate for plant-model mismatch. Starting from extreme initial guesses for the unknown growth inhibition constant, the framework converges to the optimal solution within 4 batches.