An enhanced flux balance analysis based on the state-of-the-art metaheuristics for optimizing succinate production

Flux balance analysis (FBA) is a popular algorithm in in silico metabolic engineering. This algorithm finds the optimal flux reaction under a particular perturbation, such as reaction knockout. Despite its advantages, FBA is unable to choose the best perturbation. Therefore, this study proposed an enhancement of FBA using state-of-the-art metaheuristics. Three algorithms were proposed, namely Aquila-based FBA (ABFBA), Wolf-based FBA (WBFBA), and Komodo-based FBA (KBFBA), incorporating Aquila Optimizer, Grey Wolf Optimizer, and Komodo Mlipir Algorithm, respectively, into FBA. To identify their performance, those algorithms were compared to the DSAFBA algorithm and wild-type in optimizing succinate production utilizing the E. coli strains, i.e., E. coli core, iAF1260, and iJO1366. Simulations were conducted under both aerobic and anaerobic conditions. The results demonstrated that WBFBA consistently outperformed other algorithms, achieving the highest succinate production dominantly, where the production in one microbial increased by 0.223% and 0.015% from the wild-type and the DSAFBA results, respectively. Meanwhile, ABFBA and KBFBA showed competitive performance across the strain models. Furthermore, there was an inverse relationship between growth and production rates, indicating that optimizing succinate production often reached at the cost of the growth rate. Nevertheless, the results obtained still ensured the viability of the mutant strains. In computational time, WBFBA and ABFBA achieved efficient time performance, reducing the running time by around 2% from DSAFBA’s running time. Meanwhile, KBFBA experienced increasing computational overhead due to population size adjustments. Finally, the proposed algorithms provide a promising approach for microbial strain optimization.