Universal Capacitance Model for Real-Time Biomass in Cell Culture

Area of Science:

• Biotechnology
• Bioprocess Engineering
• Cell Culture Monitoring

Background:

• Capacitance probes offer safe and robust bioprocess control by detecting cellular capacitance.
• Existing statistical biomass models lack transferability between cell lines and process conditions, leading to significant errors (over 100%) in the declining phase.
• Model transfer issues cause delays and inefficiencies in bioprocess development.

Purpose of the Study:

• To develop a single, universal biomass estimation model for capacitance probes.
• To enable model adaptation for untested cell clones and scales using a mechanistic factor.
• To improve model transferability and reduce errors throughout the entire bioprocess, including the declining phase.

Main Methods:

• Developed a novel methodology for selecting sensitive frequencies to build a statistical biomass model.
• Implemented a universal model adaptable via a simple linear factor.
• Validated the model's performance across different cell lines and process conditions.

Main Results:

• Achieved a shared statistical model with significantly reduced errors, ranging from 9% to 38% (mean ~20%) for the entire process.
• The model demonstrated improved accuracy even during the declining phase of cell culture.
• Identified a simple linear factor correlating with model transferability, linked to cell phenotype or physiology.

Conclusions:

• A universal biomass estimation model using capacitance probes is feasible and adaptable.
• The new methodology enhances model transferability and accuracy, reducing bioprocess development time.
• The identified linear factor provides insights into the physiological basis of model transferability.