Artificial Neural Network: Optimization and Characterization of α-Amylase Production from Bacillus velezensis Species.

This article delves into the application of design of experiments methodologies for the integration of a second-order definitive screening design (DSD) and artificial neural network (ANN) to comprehensively assess and predict nine operational variables aimed at increasing the yield of α-amylase from Bacillus velezensis species. By utilizing environmentally friendly and cost-effective agro-solid substrates such as moong husk and soya bean cake, the physical and chemical parameters influencing α-amylase biosynthesis from B. velezensis species were optimized, gathering early data from experiments conducted in shake flasks through the standard one-factor-at-a-time (OFAT) technique. In the realm of response surface methodology (RSM) utilizing the DSD model, nine process variables were taken into account, including pH, temperature, carbon source, nitrogen source, KPO, MgSO, NaCl, fructose, and NaNO. Furthermore, optimization based on ANN modeling was employed to enhance the enzyme yield further. Experiments were then executed under the optimal conditions as defined by RSM and ANN to corroborate the predicted optimized enzyme activity. As a result, B. velezensis species exhibited enzyme activity of 1092.49 U/mL under the optimal process variables identified by both RSM and ANN optimization methods, which included pH 5.48, temperature (34.28°C), carbon source (4.09%), nitrogen source (2.02%), KPO (0.34%), MgSO (0.14%), NaCl (0.23%), fructose (1.54%), and NaNO (0.53%). To encapsulate, compared to the OFAT technique, where the enzyme activity was 418.25 U/mL, a 2.6-fold increase in enzyme activity was achieved by integrating DSD and ANN optimization, considering only nine significant process parameters for the proliferation of B. velezensis species and the maximization of α-amylase activity. The α-amylase enzyme from the B. velezensis species was further purified and characterized. The purification process achieved a 71.77-fold increase in specific activity, with the purified enzyme exhibiting optimal activity at pH 5.5 and 55°C. The enzyme displayed high thermal stability, with minimal activity loss up to 4°C. Kinetic analysis revealed a K of 0.85 mg/mL and a V of 250 U/mg/min. The enzyme was found to be metal-independent, with inhibition observed for certain metal ions.