This proposal aims to apply data analytic solutions to challenges encountered in biomanufacturing operations. We will develop and validate two process modeling tools that manufacturers can use to quantitatively assess the process risks associated with their choice of manufacturing model. The first tool will assist manufacturers in their selection of the appropriate manufacturing model. The second tool provides a comprehensive first-principles model of a biopharmaceutical manufacturing operation which allows the user to test in silico a variety of models and quantitate the risk incurred in each choice. We will experimentally validate performance of these tools in a batch and continuous monoclonal antibody manufacturing process in a testbed facility. Additionally, we will investigate data analytic methods to appropriately incorporate textual data sources into plant-wide operational models so that manufacturers are able to utilize all of the relevant information available to them to optimize their ability to supply quality medicines to patients. Finally, we will investigate the use of data analytic methods to better connect the manufacturing process to clinical experiential data by incorporation of external data sources generated after commercial product launch, such as adverse events, outcomes data, published research, and other textual data. This work will demonstrate how data analytics can leverage additional data sources to inform manufacturers? risk-based decision making. In addition, the tools developed under this proposal will also function as training tools for regulators, who can use them to increase their own understanding of how analytical tools used in development of the manufacturing model and control strategy affect product quality and the risk incurred through selection of an inappropriate model. At its completion this work will improve (a) the regulatory process by increasing understanding around the process of choosing manufacturing process models, (b) product quality by ensuring manufacturers have the skills to choose the appropriate tools for their application, and (c) safety and efficiency through optimization of manufacturing operations.
We will develop and validate decision support tools which can be used to a) determine the appropriate process models to use for specific processes and b) quantitate the degree of risk associated with use of the selected model for process decision-making. Additionally, we will apply AI-based data analytics to incorporate data sources, both numerical and textual, which have not traditionally been used in manufacturing models and determine whether inclusion of these additional data can generate operational models that manufacturers can use to improve their product quality; and finally we will map publicly available data on adverse events to manufacturing data and assess how these data might be used for product and process understanding. This cooperative work will improve (a) the regulatory process by increasing understanding around the process of choosing manufacturing process models, (b) product quality by ensuring manufacturers have the skills to choose the appropriate tools for their application, and (c) safety and efficiency through optimization of manufacturing operations.
