Modeling and Optimization of L-Asparaginase production from novel Bacillus stratosphericus by soft computing techniques

Feed forward Artificial Neural Network (ANN) model and global optimization by Genetic Algorithm (GA) were employed on significant variables to enhance the production of L-Asparaginase from Bacillus stratosphericus. Using the experimental data, network was built with 4 inputs and 2 outputs along with 10 hidden neurons. Levenberg-Marquardt back propagation algorithm was employed to study the interactions between variables and their influence on L-Asparaginase and L-Glutaminase activity. The predicted enzyme activities were compared with the experimental data. The R² value was found to be 0.99419 from ANN and it was higher compared to Response Surface Methodology (RSM). GA optimization was employed on the quadratic equation obtained from RSM studies to find optimal solution. The optimal concentrations of L-Asparaginase and L-Glutaminase obtained from GA were 29.68 IU/ml and 0.12 IU/ml respectively. The optimal process variables were found to be Incubation time-55h, pH-6.0 and Temperature24°C and L-Asparagine-2.5 g/L.