Modeling and optimal control of integrated continuous monoclonal antibody manufacturing processes

基于单克隆抗体的药物在治疗和市场潜力方面在生物制药工业中发挥着不可或缺的作用。在治疗和诊断应用中，mAb广泛用于治疗自身免疫性疾病和癌症等。根据最近的出版物，mAb在COVID-19的治疗中也显示出有希望的结果。截至2020年9月22日，美国食品药品监督管理局（FDA）已批准了94种治疗性mAb，2010-2020年批准的mAb数量是2010年之前批准的mAb数量的三倍。就其市值而言，预计2023年将达到1982亿美元。因此，由于mAb市场的爆炸性增长，增加mAb生产工艺的生产能力是立即必要的。加拿大是开发高产能mAb生产工艺的积极和有竞争力的贡献者。mAb的集成连续生产代表了mAb生产的最新技术水平，并吸引了大量关注。然而，没有现有的集成制造过程的数学模型，也没有整个集成过程的最优控制算法。该项目将通过首先开发单克隆抗体集成连续生产过程的数学模型来填补知识空白。开发的模型将使用文献中的可用数据以及我们的工业合作伙伴的客户提供的潜在数据进行验证。基于该模型，将对基于机器学习的最优控制算法进行训练、评估和改进，以准备将其应用于工业mAb生产过程。该项目将（a）进一步提高加拿大在生物制药行业的竞争力，（B）培养具有生物制药和控制工程跨学科背景的高素质人才（HQP），以及（c）在不久的将来可能在合作伙伴组织中创造更多的就业机会。

Drugs based on mAbs play an indispensable role in biopharmaceutical industry in aspects of therapeutic and market potentials. In therapy and diagnosis applications, mAbs are widely used for the treatment of autoimmune diseases, and cancer, etc. According to a recent publication, mAbs also show promising results in the treatment of COVID-19. Until September 22, 2020, 94 therapeutic mAbs have been approved by U.S. Food & Drug Administration (FDA) and the number of mAbs approved within 2010-2020 is three times more than those approved before 2010. In terms of its market value, it is expected to reach a value of $198.2 billion in 2023. Thus, increasing the production capacity of mAb manufacturing processes is immediately necessary due to the explosive growth in mAb market. Canada is an active and competitive contributor to the development of high capacity mAb manufacturing processes. Integrated continuous manufacturing of mAbs represents the state-of-the-art in mAb manufacturing and has attracted a lot of attention. However, there is no existing mathematical model of the integrated manufacturing process and there is no optimal control algorithm of the entire integrated process. This project will fill the knowledge gaps by first developing a mathematical model of the integrated continuous manufacturing process of mAbs. The developed model will be validated using available data in the literature as well as potential data available from our industrial partner's clients. Based on the model, the machine-learning-based optimal control algorithm will be trained, assessed, and refined to get it ready for its implementation to industrial mAb manufacturing processes. This project will (a) further enhance the competitiveness of Canada in biopharmaceutical industry, (b) train highly qualified personnel (HQP) that have interdisciplinary backgrounds in biopharmaceutical and control engineering, and (c) may create more job opportunities in the partner organization in the near future.