Halfway to Automated Feeding of Chinese Hamster Ovary Cells.

Abstract in brief

• This paper discusses the development of models and soft sensors using Raman spectroscopy and chemometric techniques to automate feeding in bioreactors for Chinese hamster ovary (CHO) cells, addressing challenges like high data dimensionality and outliers.
• It highlights the importance of data preprocessing and understanding variable relationships for creating accurate models, and describes a simulation environment based on CHO cell kinetics that helps optimize nutrient dosing and analyze growth effects.
• The models were validated with high robustness and accuracy, achieving a 40% reduction in error compared to older methods, contributing to advancements in monitoring and automation in bioprocessing technology.]