The Use of Design of Experiments to Evaluate an Escherichia coli strain Under Different Fermentation Process Conditions

DoE is a useful discipline to apply in fermentation process development, however, it requires a reliable measurement systems that are often difficult to achieve in complex upstream processes.  Here we will discuss a recent upstream process improvement project made possible by a combining a rigorous DoE design with robust method development.  A Definitive Screening Design (DSD) was employed to evaluate the cell growth and product concentration for an E. coli strain under varying fermentation conditions. The experiments were conducted in 5L Sartorius A+ Bioreactors with controls for Temperature, pH, and DO levels. The fermentation process parameters evaluated with the definitive screening design were Temperature, pH, DO level, and Feed/Phosphate ratio. The responses measured for each experimental trial were OD₆₀₀ (multiple time points) Dry Cell Weight (DCW at harvest), and product concentration (at harvest). The JMP statistical software (SAS Institute, Cary, NC) was used to analyze the experimental results and estimate responses from predictive models. A standardized model selection workflow was applied to obtain the best fit models for each of the responses.  The models were subsequently used for optimization and sensitivity analysis to determine the relative impact of each experimental factor on the responses. In a final step, a confirmatory fermentation campaign was performed to verify the process models.  This revealed a surprising analytical gap around the titer method which was previously thought to be robust.