Model-based and Experimental Optimization Using Genetic Algorithm

Z. Findrik Blažević, B. Zelić, S. Bogdan and Đ. Vasić-Rački

Abstract
For the purpose of this paper genetic algorithm (GA) was used for a model-based optimization of the enzymatic synthesis of N-acetyl neuraminic acid and experimental optimization of the initial conditions of an aqueous two-phase extraction system. Model-based optimization of enzymatic synthesis of N-acetyl neuraminic acid was carried out by using an earlier developed and verified mathematical model for this system in a batch reactor.27 The influence of pyruvate (cPyr) and N-acetyl glucosamine (cGlcNAc) initial concentrations, epimerase (Eepi) and aldolase (Eald) concentrations and reaction time (t) were investigated to optimize volumetric productivity (QP) and ratio (R) of the final concentrations of N-acetyl neuraminic acid and pyruvate. In this example parameters were optimized within 40 generations of 8 experiments. Maximal volumetric productivity of QP = 47.8 g/(dm3 d) in the batch reactor and ratio of R = 2.57 were estimated under the following initial operating conditions: cPyr = 1 mmol/dm3, cGlcNAc = 460 mmol/dm3, Eepi = 39.9 mg/cm3, Eald = 0.2 mg/cm3 and t = 156.8 min. Conversion of N-acetylglucosamine to N-acetylneuraminic acid at these conditions was 0.15 %, and conversion of pyruvate to N-acetylneuraminic acid was 71.99 %. Experimental optimization using GA was applied to investigate the influence of the mass fraction of PEG-6000 (wPEG), the mass fraction of ammonium sulfate (wAS), pH and BS Albumine (BSA) concentration cBSA on partition coefficient of BSA KBSA in aqueous two-phase system consisting of PEG-6000 and ammonium sulfate. The partition coefficient was optimized within five sets of four experiments. It was found out that the mass fraction of PEG-6000 of 0.1770, the mass fraction of ammonium sulfate of 0.1105, pH of 7.0 and BSA concentration of 2.5 g/dm3 are the best working conditions to achieve the optimal partition coefficient of KBSA = 0.4793.

This work is licensed under a Creative Commons Attribution 4.0 International License

Keywords
genetic algorithm, model-based optimization, experimental optimization, enzymatic synthesis of N-acetyl neuraminic acid, aqueous two-phase system, BS Albumine