Application of machine learning methods to pathogen safety evaluation in biological manufacturing processes

Area of Science:

• Biopharmaceutical Manufacturing
• Process Development
• Machine Learning Applications

Background:

• Recombinant therapeutic protein production carries risks of viral contamination from cell lines and raw materials.
• Ensuring adequate viral clearance is critical for controlling potential contamination in biomanufacturing.
• Current viral clearance methods, like low pH inactivation, require extensive laboratory characterization for new products.

Purpose of the Study:

• To evaluate the utility of machine learning (ML) for process understanding and predictive modeling of viral clearance.
• To explore ML's potential in streamlining the development and optimization of viral clearance unit operations.
• To apply ML to a case study focusing on low pH viral inactivation for therapeutic antibody production.

Main Methods:

• Utilized machine learning techniques for process understanding.
• Developed predictive models for viral clearance.
• Conducted a case study on low pH viral inactivation.

Main Results:

• Demonstrated the potential of machine learning in modeling viral clearance.
• Showcased ML's capability to aid in process characterization for viral inactivation.
• Identified ML as a valuable tool for optimizing viral clearance unit operations.

Conclusions:

• Machine learning methods show promise for enhancing process understanding and predictive capabilities in viral clearance.
• ML can significantly streamline the development and optimization of viral clearance strategies for therapeutic antibodies.
• This approach can reduce the resources and time required for process characterization studies.