Moving towards an era of hybrid modelling: advantages and challenges of coupling mechanistic and data-driven models for upstream pharmaceutical bioprocesses
Apostolos Tsopanoglou, Ioscani Jiménez del Val
Abstract
Monoclonal antibodies (mAbs) are the highest-selling class of biopharmaceuticals due to their capabilities in treating severe illnesses. Thus, strategies to optimise their manufacture have been at the centre of recent academic and industrial research. Mechanistic and statistical modelling approaches are considered useful tools in that quest, not only because they can be employed for online process monitoring but also due to their abilities to offer useful insight into how the underlying micro and macro-scale phenomena of upstream bioprocesses impact the yield and quality of biopharmaceuticals. This manuscript provides an overview of the mechanistic and statistical models of upstream mAb bioprocesses that have been published over the past five years and discusses their advantages and drawbacks. We conclude with an outline of synergistic, hybrid modelling strategies, which are emerging as key tools in the era of Biopharma 4.0.