Comparison of viable cell concentration estimation methods for a mammalian cell cultivation process

Area of Science:

• Biotechnology
• Bioprocess Engineering
• Cell Culture Technology

Background:

• Accurate on-line estimation of viable cell concentration is crucial for optimizing fed-batch cultivation of Chinese Hamster Ovary (CHO) cells.
• Traditional off-line measurements are time-consuming and limit real-time process control.

Purpose of the Study:

• To evaluate various mechanistic and black-box models for on-line estimation of viable cell concentrations in CHO cell fed-batch cultures.
• To compare the performance of different modeling approaches using independent validation data.

Main Methods:

• Application of mechanistic, black-box, empirical, and linear models.
• Identification of models using data from six fed-batch cultivation experiments.
• Validation of model performance using six independent data sets.
• Quantification of performance using root mean square error (RMSE).

Main Results:

• Simple empirical and linear models demonstrated good on-line estimation performance.
• Hybrid models, multivariate linear regression, and support vector regression achieved the best results.
• Hybrid models provided valuable insights into specific cellular growth rates.

Conclusions:

• Multiple modeling strategies can effectively estimate viable cell concentrations in CHO cell cultures.
• Advanced techniques like hybrid models, MLR, and SVR offer superior performance for bioprocess monitoring.
• On-line estimation enhances real-time control and optimization of biomanufacturing processes.