S14: Efficient Characterization of a Biomanufacturing Fermentation Process Using Definitive Screening Designs

Development of a bio-manufacturing process requires that critical process parameters be identified early on and a suitable process model be developed to accurately predict process behavior as a function of those critical parameters.  Recently, Jones and Nachtscheim (2011) introduced a new class of highly efficient experimental designs referred to as Definitive Screening designs that can allow characterization and optimization of a bio-process with a single experiment.  In this talk we show how the use of a single 5 factor, 15 run Definitive Screening design resulted in the successful characterization and modeling of a fermentation step in a bio-manufacturing process.  The talk will feature a fermentation case study to demonstrate the power of the Definitive Screening Designs for characterization and Optimization. Also, we will show how to use modern statistical modeling tools (found in the JMP 11 statistical software) to find the best possible statistical models to predict fermentation process behavior. These models are then used for optimization and sensitivity analysis (a.k.a. range finding) of the fermentation process.