Model-assisted Design of Experiments as a concept for knowledge-based bioprocess development.

Area of Science:

• Biotechnology
• Process Engineering
• Quality by Design (QbD)

Background:

• Traditional Design of Experiments (DoE) in bioprocess development requires user-defined factor boundaries, often leading to extensive, iterative experimentation.
• This limitation can result in significant time and cost investments during process optimization.

Purpose of the Study:

• To introduce a model-assisted DoE concept for knowledge-based reduction of factor boundary values in bioprocess development.
• To demonstrate a method for simulating experimental space to optimize bioprocesses, thereby reducing the number of required physical experiments.

Main Methods:

• Estimation of parameters for a mathematical process model.
• Simulation of factor combinations using the developed model, replacing traditional experimental derivation.
• Constraint-based evaluation and optimization of the experimental space.

Main Results:

• The model-assisted DoE approach identified the same optimal process strategies as traditional DoE for an antibody-producing Chinese hamster ovary (CHO) batch and fed-batch process.
• Model-assisted DoE required only 4 experiments, compared to 16 for batch and 29 for fed-batch processes using traditional DoE.
• Significant reduction in the number of experiments needed for knowledge-based bioprocess development was achieved.

Conclusions:

• Model-assisted DoE offers a more efficient and cost-effective strategy for bioprocess development and optimization.
• This simulation-based approach effectively reduces experimental burden while maintaining the identification of optimal process parameters.
• The methodology is applicable to antibody-producing CHO cell culture processes, highlighting its practical value in the biopharmaceutical industry.