(183c) Simultaneous Process Design & Optimisation: An Application to Bioprocesses

Traditional biopharmaceutical process development relies primarily on process- and product- know-how and is aided by wet-lab experimentation. Although this may have proven to be a suitable strategy, able to identify designs that guarantee high product quality and titres, it is often cost- and time- inefficient (Narayanan et al., 2020; Nasr et al., 2017). At the same time, it does not allow for combinatorial studies to clearly identify underlying synergetic and/or antagonistic effects of the process parameters on often-competing Key Performance Indicators (KPIs), such as product purity and process yield.

In this work, we propose a framework for the simultaneous design and optimisation of process design and operation. The presented methodology is based on an experimentally validated high-fidelity process model that is used as the virtual experimentation platform (digital twin). The digital twin forms the basis for the execution of Quasi Random Sampling-High dimensional model representation experiments, where design and process parameters are varied within the range of interest. The obtained datasets are subsequently assessed based on process performance with respect to predefined KPIs. A concave hull is constructed by implementing the alpha shape algorithm to identify the Normal Operating Region (NOR) where all constraints are satisfied.

Here we showcase the capabilities of the framework on an affinity chromatography purification step, used in monoclonal antibody (mAb) manufacturing. We use the high-fidelity process model as presented by Steinebach et al. (2016) and we focus on assessing the performance of the system: (a) using and comparing three industrially relevant resins (MabSelect SuRe, MabSelect SuRe LX, and CaptivA PriMAB) and (b) under measured disturbances resulting from variations in the feed composition. As KPIs of interest, we focus on process yield and productivity and we use those for the design of the NOR in each case.

We demonstrate how our proposed framework can be used to calculate the size of the NOR and quantify a flexibility region (FR) around a chosen operating point. The FR is translated into the allowed deviation of the process conditions, in order for optimal operation to be guaranteed. A uniform maximum deviation with respect to the feed conditions, where the constraints are still satisfied can be quantified for the different resins. The maximum value of productivity inside the NOR is also quantified. Currently, the framework is being expanded to different unit operations both in bioprocesses and beyond, such as energy systems.

References

Narayanan, H., Luna, M. F., von Stosch, M., Cruz Bournazou, M. N., Polotti, G., Morbidelli, M., Butte, A. & Sokolov, M. (2020) Bioprocessing in the Digital Age: The Role of Process Models. Biotechnol J. 15 (1), e1900172.

Nasr, M. M., Krumme, M., Matsuda, Y., Trout, B. L., Badman, C., Mascia, S., Cooney, C. L., Jensen, K. D., Florence, A., Johnston, C., Konstantinov, K. & Lee, S. L. (2017) Regulatory Perspectives on Continuous Pharmaceutical Manufacturing: Moving From Theory to Practice: September 26-27, 2016, International Symposium on the Continuous Manufacturing of Pharmaceuticals. J Pharm Sci. 106 (11), 3199-3206.

Steinebach, F., Angarita, M., Karst, D. J., Muller-Spath, T. & Morbidelli, M. (2016) Model based adaptive control of a continuous capture process for monoclonal antibodies production. J Chromatogr A. 1444, 50-56.