Evaluating Simulated Annealing Algorithms in the Optimization of Bacterial Strains

In this work, a Simulated Annealing (SA) algorithm is proposed for a Metabolic Engineering task: the optimization of the set of gene deletions to apply to a microbial strain to achieve a desired production goal. Each mutant strain is evaluated by simulating its phenotype using the Flux-Balance Analysis approach, under the premise that microorganisms have maximized their growth along natural evolution. A set based representation is used in the SA to encode variable sized solutions, enabling the automatic discovery of the ideal number of gene deletions. The approach was compared to the use of Evolutionary Algorithms (EAs) to solve the same task. Two case studies are presented considering the production of succinic and lactic acid as the target, with the bacterium E. coli. The variable sized SA seems to be the best alternative, outperforming the EAs, showing a fast convergence and low variability among the several runs and also enabing the automatic discovery of the ideal number of knockouts.