TY - JOUR
T1 - Modeling and optimization of tannase production with Triphala in packed bed reactor by response surface methodology, genetic algorithm, and artificial neural network
AU - Selvaraj, Subbalaxmi
AU - Vytla, Ramachandra Murty
AU - Vijay, G. S.
AU - Natarajan, Kannan
N1 - Publisher Copyright:
© 2019, King Abdulaziz City for Science and Technology.
PY - 2019/7/1
Y1 - 2019/7/1
N2 - In this research, optimization of the production medium to enhance tannase production by Bacillus gottheilii M2S2 in laboratory-scale packed bed reactor was studied. Amount of substrate Triphala, moisture content, aeration rate, and fermentation period was chosen for optimization study. During one variable at a time optimization, the highest tannase activity of 0.226 U/gds was shown with Triphala as a substrate at the fermentation period of 32 h. Furthermore, the optimum conditions predicted by response surface methodology (RSM) and genetic algorithm (GA) were found to be 11.532 g of substrate Triphala, 47.071% of the moisture content, and 1.188 L/min of an aeration rate with uppermost tannase activity of 0.262 U/gds. In addition, the single hidden layer feedforward neural network (SLFNN) and the radial basis function neural network (RBFNN) of an artificial neural network (ANN) were adopted to compare the prediction performances of the RSM and GA. It revealed that the ANN models (SLFNN, R2 = 0.9930; and RBFNN, R2 = 0.9949) were better predictors than the RSM (R2 = 0.9864). Finally, the validation experiment exhibited 0.265 U/gds of tannase activity at the optimized conditions, which is an 11-fold increase compared to unoptimized media in shake flask.
AB - In this research, optimization of the production medium to enhance tannase production by Bacillus gottheilii M2S2 in laboratory-scale packed bed reactor was studied. Amount of substrate Triphala, moisture content, aeration rate, and fermentation period was chosen for optimization study. During one variable at a time optimization, the highest tannase activity of 0.226 U/gds was shown with Triphala as a substrate at the fermentation period of 32 h. Furthermore, the optimum conditions predicted by response surface methodology (RSM) and genetic algorithm (GA) were found to be 11.532 g of substrate Triphala, 47.071% of the moisture content, and 1.188 L/min of an aeration rate with uppermost tannase activity of 0.262 U/gds. In addition, the single hidden layer feedforward neural network (SLFNN) and the radial basis function neural network (RBFNN) of an artificial neural network (ANN) were adopted to compare the prediction performances of the RSM and GA. It revealed that the ANN models (SLFNN, R2 = 0.9930; and RBFNN, R2 = 0.9949) were better predictors than the RSM (R2 = 0.9864). Finally, the validation experiment exhibited 0.265 U/gds of tannase activity at the optimized conditions, which is an 11-fold increase compared to unoptimized media in shake flask.
UR - http://www.scopus.com/inward/record.url?scp=85067066234&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85067066234&partnerID=8YFLogxK
U2 - 10.1007/s13205-019-1763-z
DO - 10.1007/s13205-019-1763-z
M3 - Article
AN - SCOPUS:85067066234
SN - 2190-572X
VL - 9
JO - 3 Biotech
JF - 3 Biotech
IS - 7
M1 - 259
ER -