DATA-BASE SIMULATION OF BATCH AND FED-BATCH CULTURES FOR ALPHA-AMYLASE PRODUCTION USING A CULTURE DATA-BASE AND A STATISTICAL PROCEDURE

A practical simulation method using a culture data base and a statistical procedure was developed. This Data Base Simulation (DBS) method enables various time courses of batch and fed-batch cultures to be simulated by using multiple regression models in combination with a culture data base. The culture data base contained culture data and specific rates (mu, etc.), which were smoothed and estimated using a cubic spline function based on raw data gained under various experimental conditions. The cubic spline function was effective in improving data smoothing, data interpolation and the estimation of specific rates from raw culture data. The coefficients of the regression models were estimated by regression analysis using data sets statistically selected from the data base, followed by numerical calculation of the time courses of state variables over specified intervals. Then the coefficients were renewed and the simulation was continued to the end of the cultivation by repeating the coefficient renewals. The method was tested using experimental data from batch and fed-batch cultures for alpha-amylase production by Bacillus amyloliquefaciens. The culture data base was successfully produced by applying a cubic spline function to more than 30 pieces of experimental data. The simulated results showed that the method was able to appropriately predict the time-courses of non-linear batch and fed-batch cultures for a wide range of culture conditions without using specific deterministic models.