🤝 Collaborating with AstraZeneca, this model helps screen drug candidates early, saving costs in formulation.
🔒 It addresses data sharing challenges while maintaining confidentiality, achieving 90% accuracy using training data without disclosing intellectual property.
🌐 This model could revolutionize drug development in biomanufacturing by streamlining processes.
Introduction:
The article discusses a significant advancement in deep learning applications within biomanufacturing, particularly concerning the production of monoclonal antibodies (mAbs). Researchers from the Stevens Institute of Technology developed a machine learning model in collaboration with AstraZeneca to predict the viscosity of mAb formulations, facilitating early screening of drug candidates while ensuring the confidentiality of proprietary data.
- Researchers have developed a machine learning model that estimates the viscosity of monoclonal antibody formulations to streamline drug development.
- AstraZeneca shared specific mAb sequences and viscosity data with researchers under a confidentiality agreement, enabling model training without compromising intellectual property.
- The model predicts which mAbs may exhibit high viscosity during subcutaneous injection, allowing for proactive candidate screening.
- Initial tests demonstrated the model’s accuracy of nearly 90%, bolstering confidence in its future applications across the industry.
- Similar collaborative projects are enabling companies to contribute data while protecting sensitive information, showcasing a new approach to data sharing in biomanufacturing.
Conclusion:
The development of this deep learning model signifies a pivotal stride in biomanufacturing, allowing for efficient candidate screening while maintaining data confidentiality. This approach not only addresses challenges related to data sharing in drug development but also sets the stage for future innovations in the field, ultimately aiming to enhance the efficiency and effectiveness of monoclonal antibody production processes.
