(442c) Enhancing Mab Production through Hybrid Modeling for Real-Time Process Control

The production of monoclonal antibodies (mAbs) plays a critical role in biopharmaceutical therapies, driving the need for efficient and scalable manufacturing processes. Despite advancements in automation technologies, bioreactor operations still often rely on input from subject matter experts (SMEs) for real-time decision-making based on batch progression. However, decisions based solely on SME empirical knowledge can result in inconsistencies between batches and significant time investment.

This study introduces a model-based approach for real-time predictions of cell culture dynamics and critical quality attributes (CQAs), enabling SMEs to make data-informed decisions to optimize process control, thereby improving product yield and quality consistency. Specifically, we evaluated the implementation of a hybrid model that integrates mechanistic frameworks with machine learning algorithms.

This presentation will highlight key steps in model development, demonstrate practical applications with specific use cases currently implemented in manufacturing, and present a vision for the future of hybrid modeling in bioproduction. Our findings showcase the potential of hybrid models to enhance operations in Good Manufacturing Practice (GMP) environments, paving the way for more efficient, consistent, and sustainable production. Furthermore, this approach is adaptable to a wide range of cell lines and biomanufacturing products.