Data-based optimization of protein production processes.

Area of Science:

• Biochemical Engineering
• Process Optimization
• Biotechnology

Background:

• Data-based modeling is established, but data-based optimization methods are underdeveloped.
• Existing methods often rely on complex models rather than direct experimental data.
• Dynamic programming principles offer a potential framework for data-driven optimization.

Purpose of the Study:

• To present a novel data-based optimization technique for bioprocesses.
• To utilize direct experimental data for optimizing adjustable variables.
• To improve fermentation runs through optimized control profiles.

Main Methods:

• Developed a data-based optimization technique founded on dynamic programming principles.
• Formulated the performance index J as a function of biomass (x) instead of time (t).
• Employed mathematical programming to derive optimal control paths u(opt)(x) from J(x) derivatives.

Main Results:

• Demonstrated the feasibility of the method through numerical experiments.
• Successfully applied the optimization technique to recombinant protein formation in E. coli.
• Presented experimental validation results confirming the method's effectiveness.

Conclusions:

• The proposed data-based optimization technique is feasible and effective for bioprocesses.
• Reformulating performance as a function of biomass simplifies optimization using adjustable variables.
• This method allows for the derivation of optimal control profiles directly from experimental data, leading to improved process outcomes.