Novel data science Methods as enablers for robust Control strategies in monoclonal antibody production

While other industries managed to break the six-sigma barrier, the pharmaceutical industry remains a3 sigma industry when it comes to biologics production.Chemical processes are mostly well understood, while in biological process a lot of the variation still remains unexplained. However, understanding this variation is the key of rendering processes more robust. Setting up holistic control strategies is one of the four pillars of Pharma 4.0. The topic of industry 4.0 emerged in 2011. Pharma 4.0 aims to transform pharmaceutical manufacturing by applying digital strategies. Leveraging data and applying novel statistical tools are key to bring an ICHholistic control strategy to life and thereby minimizing the manufacturing batch failure rate.Within this thesis data science methods are applied to overcome obstacles and fill knowledge gaps onthe way of the pharmaceutical industry to a six sigma industry with focus on process development and validation for the manufacturing of biopharmaceuticals.Integrated process modelling approaches are used of for setting up proven acceptable ranges and setting up acceptance limits. This allows to set up control strategies based on out of specification probabilities and thereby providing a novel tool for managing the risk when setting up controlstrategies. A method for criticality assessment by means of retrospective power analysis is presented that minimizes the risk of overlooking potentially critical effects in process validation. In addition, to that a novel approach to temperature control in mAb production in mammalian cells was investigated and applied in manufacturing scale.The techniques developed and applied in this thesis enable pharmaceutical manufacturer to set upmore robust control strategies without the need of performing additional wet-lab experiments, just by exploiting available data more efficiently. These methodologies put the regulatory guidelines into practice and provide a solid line of reasoning for market authorization. In addition to the data sciencemethods, an of-the-beaten-track cultivation strategy for a mAb production in mammalian cell culture ispresented that might open up novel degrees of freedom for manufacturers steering the product into biosimilarity.