Optimization of polyhydroxybutyrate (PHB) production by Azohydromonas lata MTCC 2311 by using genetic algorithm based on artificial neural network and response surface methodology

Abstract

Abstract In the present study, the maximum biomass and polyhydroxybutyrate productions were studied and optimized using suitable carbon and nitrogen sources by bacterial strain Azohydromonas lata MTCC 2311. Among three carbon sources namely, sucrose, fructose, and glucose and four nitrogen sources namely, (NH4)2SO4, NH4Cl, urea, and NH4NO3 studied in shake flask experiments, sucrose and urea were found to be the best carbon and nitrogen sources, respectively. Further, response surface methodology (RSM) and artificial neural network models (ANN) were applied to navigate the experimental data obtained in accordance with the central composite design. The effects of sucrose (3.2–36.82 g/L), urea (0.16–1.84 g/L), and TE solution (0.32–3.68 ml/L) on biomass and PHB concentrations were investigated. The modeling and optimization ability of hybrid ANN-GA had shown higher accuracy in finding optimum concentrations of medium variables than hybrid RSM-GA. Hybrid ANN-GA predicted the maximum biomass concentration (12.25 g/L) at the optimum level of medium variables: sucrose, 35.27 g/L; urea, 1.55 g/L; and TE solution, 0.42 ml/L. Whereas, the maximum predicted PHB concentration (5.95 g/L) was reported at: sucrose, 35.20 g/L; urea, 1.58 g/L; and TE solution, 0.36 ml/L. The validation with additional set of data shows that the predictive errors (%) in biomass and PHB concentrations were 3.67 and 2.52, respectively for shake flask experiments, whereas, the predictive errors (%) were 13.80 and 14.28, respectively, for bioreactor experiments.