Application of Artificial Neural Network Technique for the Production of Biotoxin

Abstract

In this study. a locally isolated strain of Bacillus thuringiensis that has insecticidal activity against dengue vector (larvae of Aedes aegypti), was cultivated. Different carbon and nitrogen sources were screened for enhanced bacterial growth. The factors affecting Bacillus thuringiensis’s biomass production like concentration of carbon, nitrogen, pH and temperature were optimized by one parameter at a time technique. The optimal levels of the selected parameters were also obtained by using an Artificial Neural Network (ANN). Peptone and molasses were selected as the best nitrogen and carbon sources respectively. The optimal levels obtained for nitrogen, carbon, pH and temperature by using the one parameter at a time technique were 1%, 0.25%, 8, and 37 ℃ respectively with 0.53 mg/mL biomass production. The ANN predicted levels were 1% for nitrogen, 0.25% for carbon, 9 pH and 31 ℃ for temperature with the predicted value of biomass being 0.85 mg/ml. The biomass produced at predicted optimum levels of variables was 0.82 mg/ml, very close to the predicted value of 0.85 mg/ml.