Machine Learning Enables Real-Time Bioprocess Optimization
An integrated machine learning framework adapts culture conditions during long perfusion runs
-
1
Researchers at Merck developed a self-driving bioprocessing platform that optimizes monoclonal antibody production using machine learning.
-
2
The platform autonomously designs and adapts perfusion cell culture experiments in real time, enhancing decision-making during cultivation.
-
3
It integrates Bayesian experimental design with a cognitive digital twin, allowing continuous learning and real-time adjustments.
-
4
During a 27-day monoclonal antibody production run, the system autonomously operated for 20 days, meeting performance targets.
-
5
This innovative approach reduces the need for manual intervention and could shorten development timelines in biopharmaceutical production.
