AI model promises faster monoclonal antibody production
Faster clone selection could lower development costs for antibody-based medicines.
Researchers at the University of Oklahoma have developed a machine-learning model that could significantly speed up the manufacturing of monoclonal antibodies, a fast-growing class of therapies used to treat cancer, autoimmune disorders, and other diseases.
The study, published in Communications Engineering, targets delays in selecting high-performing cell lines during antibody production. Output varies widely between Chinese hamster ovary cell clones, forcing manufacturers to spend weeks screening for high yields.
By analysing early growth data, the researchers trained a model to predict antibody productivity far earlier in the process. Using only the first 9 days of data, it forecast production trends through day 16 and identified higher-performing clones in more than 76% of tests.
The model was developed with Oklahoma-based contract manufacturer Wheeler Bio, combining production data with established growth equations. Although further validation is needed, early results suggest shorter timelines and lower manufacturing costs.
The work forms part of a wider US-funded programme to strengthen biotechnology manufacturing capacity, highlighting how AI is being applied to practical industrial bottlenecks rather than solely to laboratory experimentation.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!