17-3 |
|
G. ZHANG1, G. Oddone2, D. A. Mills1, and D. E. Block1. (1) Dept. of Viticulture and Enology, Univ. of California, One Shields Avenue, Davis, CA 95616, (2) Dept.of Viticulture and Enology, Univ. of California, One Shields Avenue Davis, Davis, CA 95616 Probiotic lactic acid bacteria (LAB) are increasingly considered ideal vehicles for the delivery of intact bioactive antigens to animal gastrointestinal gut. However, optimization of recombinant protein production in LAB has not been thoroughly examined as with other potential host organisms. Since more than ten process parameters are typically included in optimizing fermentation processes, this optimization can be very difficult. Our objective is to improve the production of engineered probiotic LAB, as well as to develop a general method for optimization of complex processes in as few experiments as possible. To do this, a new approach which combines neural networks and stochastic optimization with data mining was developed to suggest optimal operating conditions for future experimentation. The method developed was compared to traditional statistical design of experiment (DOE) optimization methods using simulations. The simulation results demonstrated that more experimental points in a round of experimentation and a reasonable critical distance (a measure of the distance between experimental conditions examined) improved the results of the algorithm, but higher dimensionality and noise levels decreased the performance. Except for the case of critical distance, all resulted in a higher number of experiments required to find the optimum. To evaluate this methodology experimentally, we are using L. lactis producing Green Fluorescent Protein (GFP) as a model of recombinant protein, as well as the same strain producing Intimin, a proven antigen for E. coli O157:H7. Use of the optimization algorithm will be illustrated using data from 5 L bench-scale fermentations. Oral delivery of LAB-based vaccines produced through an optimized process could result in a less expensive route for reducing the spread of E. coli O157:H7 in animals and humans.
Session 17, Biotechnology: General
2005 IFT Annual Meeting, July 15-20 - New Orleans, Louisiana |