Strategy of Experimentation for Fermentation Process and Product Development

Design of experiments (DOE) is a statistical tool for simultaneously varying a number of factors rather than just one at a time.  Not only is DOE, being a parallel-testing method, more efficient than the traditional one-factor-at-a-time (OFAT) approach (serial), it also reveals interactions that OFAT, by definition, cannot estimate.  However, like any powerful tool, DOE achieves even greater impact by being employed in a strategic manner.  This presentation lays out the following tried-and-true strategy of experimentation:

• Screening the vital few factors (typically 20%) from the trivial many (80%)
• Characterizing main effects and interactions
• Optimizing via response surface methods (RSM)

This “SCO” strategy, used with great success by DuPont and others in process industries, that features a series of steps with decision points along the way that allow process/product developers to adapt based on the results at each stage.  Last, but not least, a necessary postscript to this strategic experimental protocol, is a series of runs or a final DOE to confirm the model developed via SCO.

A number of fermentation case studies will be detailed in this presentation.  It will encompass factorial, RSM and mixture designs of particular interest to biotechnologists.