Optimization Of The Fermentation Media Using Statistical Approach and Artifi cal Neural Networks for the Production of An Alkaline Protease from Bacillus subtilis.

Abstract

Optimization of the fermentation medium for maximum alkaline protease production was carried out by Bacillus subtilis using
Artifi cial Neural Networks (ANN) and Response Surface Methodology (RSM) and a predictive model was built, for the combined
effects of independent variables (moisture content, concentration of carbon and nitrogen supplementation). Maximum alkaline
protease produced was 701.9 U/ml with the application of RSM and the protease production further increased to 753.453 U/ml
when ANN was used. The results demonstrated a higher prediction accuracy of ANN when compared to RSM. The domination of
ANN over other multi factorial approaches would make this estimation technique a very helpful tool for fermentation monitoring
and control.