Hybrid approach for metabolites production using differential evolution and minimization of metabolic adjustment

Abstract

Microbial strains can be optimized using metabolic engineering which implements gene knockout techniques. These techniques manipulate potential genes to increase the yield of metabolites through restructuring metabolic networks. Nowadays, several hybrid optimization algorithms have been proposed to optimize the microbial strains. However, the existing algorithms were unable to obtain optimal strains because the nonessential genes are hardly to be diagnosed and need to be removed due to high complexity of metabolic network. Therefore, the main goal of this study is to overcome the limitation of the existing algorithms by proposing a hybrid of Differential Evolution and Minimization of Metabolic Adjustments (DEMOMA). Differential Evolution (DE) is known as population-based stochastic search algorithm with few tuneable parameter control. Minimization of Metabolic Adjustment (MOMA) is one of the constraint based algorithms which act to simulate the cellular metabolism after perturbation (gene knockout) occurred to the metabolic model. The strength of MOMA is the ability to simulate the strains that have undergone mutation precisely compared to Flux Balance Analysis. The data set used for the production of fumaric acid is S. cerevisiae whereas data set for lycopene production is Y. lipolytica metabolic networks model. Experimental results show that the DEMOMA was able to improve the growth rate for the fumaric acid production rate while for the lycopene production, Biomass Product Coupled Yield (BPCY) and production rate were both able to be optimized.