Modeling and optimization of fermentation variables for enhanced production of lactase by isolated Bacillus subtilis strain VUVD001 using artificial neural networking and response surface methodology.

Abstract

Modeling and optimization were performed to enhance production of lactase through submerged fermentation by Bacillus subtilis VUVD001 using artificial neural networks (ANN) and response surface methodology (RSM). The effect of process parameters namely temperature (°C), pH, and incubation time (h) and their combinational interactions on production was studied in shake flask culture by Box-Behnken design. The model was validated by conducting an experiment at optimized process variables which gave the maximum lactase activity of 91.32U/ml. Compared to traditional activity, 3.48-folds improved production was obtained after RSM optimization. This study clearly shows that both RSM and ANN models provided desired predictions. However, compared with RSM (R 2=0.9496), the ANN model (R 2=0.99456) gave a better prediction for the production of lactase.