(413e) Process Scale-up – the Weakest Link in Realization of the Promising Potential of Biotechnology

Biotechnology is set to become a formidable manufacturing platform thanks to recent advances, decades in making, in sequencing DNA (omics), gene editing (CRISPR), computing, automation, and artificial intelligence (AI). This has increased ability to understand and engineer biology of microorganism (hence host organisms), which are able to express many useful molecules at industrially relevant titers. As a result, fermentation emerged as a technology platform to produce a variety of molecules from energy, chemicals and healthcare/therapeutics to food, textile (such as artificial silk), polymers/coatings. While the molecule discovery side is looking bright, often the process scale-up side is underestimated. Indeed design and validation of a cost-effective process in a reasonable amount of time/resources (no one has infinite time and resources) is one of the key factor that determines a failed biotech innovation from a successful one. This biotech process-scale up problem present exciting set of opportunities for process systems engineering (PSE) community but equally many challenges to work on. In this talk, I discuss our experiences in application of PSE including model-based & data-based methods to various biotech manufacturing studies in the past decade. The key challenge we identify is that while PSE systems thinking is useful, however application of PSE tools and methods (design, simulation, control, optimization, planning and scheduling, capacity analysis,...) themselves require time and resources (no free lunch theorem). The main resources involved are data collection at pilot/lab and full-scale experiments and measurements. The data needs are often unique for different host organisms used to express the product of interest in fermentation reactors. Hence to benefit from PSE methods, which have been time tested and validated across many process industries (chemicals, polymers, fuels,...), the experimentation research carried for process characterization needs should be planned jointly and aligned with the simulation/PSE needs. To make a real difference, in my view, the future PSE professionals should embody an interdisciplinary expertise going beyond working only in virtual/simulation experiments to wet/real life experiments. PSE need to also address the efficient use and maintenance of models and data flow in the process operation stage (life cycle of the process), which are not trivial.Last, the biotech community should stop thinking the job is done with the discovery of the molecule! In fact, it is just the beginning of a potentially long and high risk/high reward journey, where failure is one of the outcomes especially when a new host organism is selected for which prior experience at production scale is absent. Process scale up is a hard problem. All the relevant disciplines and stake-holders from fermentation science to downstream separation science and technology, from biology to PSE (computing/simulation technologies) and ML are needed to work together to make this works in this new era of biotechnology.