![]() |
JOURNAL TOOLS |
eTOC |
To subscribe |
Submit an article |
Recommend to your librarian |
ARTICLE TOOLS |
Reprints |
Permissions |

YOUR ACCOUNT
YOUR ORDERS
SHOPPING BASKET
Items: 0
Total amount: € 0,00
HOW TO ORDER
YOUR SUBSCRIPTIONS
YOUR ARTICLES
YOUR EBOOKS
COUPON
ACCESSIBILITY
ORIGINAL ARTICLES
Minerva Biotecnologica 2010 September-December;22(3-4):63-73
Copyright © 2011 EDIZIONI MINERVA MEDICA
language: English
Comparison of predictive capabilities of response surface methodology and artificial neural network for optimization of periplasmic interferon-a2b production by recombinant Escherichia coli
Tan J. S. 1, Ramanan R. N. 2, Ling T. C. 3, Shuhaimi M. 4, Ariff A. B. 1-5
1 Unit of Immunotherapeutic and Vaccine, Laboratory of Molecular Biomedicine, Institute of Bioscience, Universiti Putra Malaysia Serdang, Selangor, Malaysia 2 Chemical and Sustainable Process Engineering Research Group, School of Engineering, Monash University, Bandar Sunway, Malaysia 3 Department of Process and Food Engineering, Faculty of Engineering, Universiti Putra Malaysia, Serdang, Selangor, Malaysia 4 Department of Microbiology, Faculty of Biotechnology and Biomolecular Sciences, Universiti Putra Malaysia Serdang, Selangor, Malaysia 5 Department of Bioprocess Technology, Faculty of Biotechnology and Biomolecular Sciences Universiti Putra Malaysia, Serdang, Selangor, Malaysia
Aim. Predictive capability of response surface methodology (RSM) and artificial neural network (ANN) in modeling growth of recombinant Escherichia coli (E. coli) and production of periplasmic interferon-α2b (PrIFN)-α2b was compared.
Methods. The effect of potassium phosphate buffer at different pH and strengths, and EDTA concentration against the production of PrIFN-α2b by E. coli were simulated using RSM and ANN. A central composite design (CCD) and multilayered feed-forward (MFF) incremental propagation algorithm of ANN were developed to predict the three variables.
Results. In the validation results, the maximum cell growth (5.52 g/L) and production of PrIFN-α2b (121.4 ng/mL) using RSM model were obtained at pH 7.13, 0.1 M phosphate buffer and EDTA concentration of 0.047 mM. On the other hand, the optimized condition derived by ANN model (dry cell weight of 5.43 g/L and PrIFN-α2b production of 130.5 ng/mL) was obtained at pH 7.00, 0.124 M phosphate buffer and EDTA concentration of 0.07 mM.
Conclusion. The results indicated that ANN had better accuracy and predictive capability as compared to RSM for the nonlinear behavior of data sets. The final optimized medium was 30% higher as compared to non-optimized phosphate buffer medium in shake flask culture.