Design of Experiments and Artificial Neural Network Linked Genetic Algorithm for Modeling and Optimization of L-asparaginase Production by Aspergillus terreus MTCC 1782

摘要

The sequential optimization strategy for design of an experimental and artificial neural network (ANN) linked genetic algorithm (GA) were applied to evaluate and optimize media component for L-asparaginase production by Aspergillus terreus MTCC 1782 in submerged fermentation. The significant media components identified by Plackett-Burman design (PBD) were fitted into a second order polynomial model (R~2 = 0.910) and optimized for maximum L-asparaginase production using a five-level central composite design (CCD). A nonlinear model describing the effect of variables on L-asparaginase production was developed (R~2 = 0.995) and optimized by a back propagation NN linked GA. Ground nut oil cake (GNOC) flour 3.99% (w/v), sodium nitrate (NaNO_3) 1.04%, L-asparagine 1.84%, and sucrose 0.64% were found to be the optimum concentration with a maximum predicted L-asparaginase activity of 36.64 IU/mL using a back propagation NN linked GA. The experimental activity of 36.97 IU/mL obtained using the optimum concentration of media components is close to the predicted L-asparaginase activity of the ANN linked GA.