Online optimal experimental re-design in robotic parallel fed-batch cultivation facilities

Area of Science:

• Biotechnology
• Bioprocess Engineering
• Mathematical Modeling

Background:

• Model identification in biosciences is time-consuming, hindering the application of mathematical modeling.
• Accurate parameter estimation for macro-kinetic growth models is crucial for bioprocess development.
• Current methods for experimental design are often inefficient, requiring significant experimental effort.

Purpose of the Study:

• To develop an integrated framework for online optimal experimental re-design in parallel nonlinear dynamic processes.
• To precisely estimate parameters of macro-kinetic growth models with minimal experimental effort.
• To enable rapid validation of models for new strains, mutants, or products in bioprocesses.

Main Methods:

• Implementation of a framework for online optimal experimental re-design.
• Utilizing automated liquid handling robots and mini-bioreactors for parallel cultivation.
• Automated at-line analyses integrated with a modeling environment for continuous data utilization.
• Online re-computation of optimal experiments based on periodical parameter estimations.

Main Results:

• Demonstrated fitting of a macro-kinetic differential equation model for Escherichia coli fed-batch processes within 6 hours.
• Achieved a 50-fold reduction in the average coefficient of variation for parameter estimates compared to sequential methods (4.83% vs. 235.86%).
• Successfully validated the framework in a complex system involving automated robots and real-time data analysis.

Conclusions:

• The developed online optimal experimental re-design framework significantly enhances the efficiency of parameter estimation for macro-kinetic models.
• This approach accelerates computer-aided bioprocess development by enabling rapid model validation and refinement.
• The system offers a systematic solution for efficient model identification, crucial for advancing biosciences.