Related Experiment Videos
Chemometric modelling based on 2D-fluorescence spectra without a calibration measurement
1Institut für Technische Chemie, University of Hannover, Callinstr. 3, 30559 Hannover, Germany.
Bioinformatics (Oxford, England)
|January 23, 2003
Summary
This study introduces a new method for calibrating chemometric models used in bioprocess supervision, eliminating the need for extensive off-line measurements. This approach leverages a priori process knowledge and mathematical modeling for accurate predictions in real-time bioprocess monitoring.
Area of Science:
- Biotechnology
- Analytical Chemistry
- Process Engineering
Background:
- 2D fluorescence spectra offer insights into intracellular compounds like NADH and flavins.
- Chemometric modeling is crucial for bioprocess supervision and control, typically requiring extensive off-line calibration data.
- Current methods necessitate numerous off-line measurements for accurate chemometric model calibration.
Purpose of the Study:
- To present a novel methodology for chemometric model calibration.
- To enable calibration without requiring additional off-line measurements.
- To improve the efficiency of bioprocess monitoring and control.
Main Methods:
- Utilizes a priori knowledge of the process, including mathematical models.
- Estimates model parameters during the calibration procedure.
- Incorporates the knowledge of substrate consumption at the end of a process run.
Main Results:
- A new chemometric calibration methodology is presented.
- The method was successfully applied to a batch cultivation of S. cerevisiae.
- Validated chemometric models derived from this method can be used for predictions in new process runs.
Conclusions:
- The developed methodology significantly reduces the need for off-line measurements in chemometric calibration.
- This approach enhances the practicality and efficiency of real-time bioprocess supervision.
- The validated models are suitable for predictive applications in bioprocesses.