No abstract

As part of the process-characterization campaign of a candidate vaccine product, a recently developed class of three-level designs-definitive-screening designs-was employed to select a quadratic model that describes the effect of six input process parameters, including protein concentration, formaldehyde-to-protein ratio, lysine concentration, reaction duration, pH, and reaction temperature, on a formylation protein-crosslinking reaction. This design requires only 17 experimental runs. The resulting model was then used to simulate 10,000 runs that account for the variability in the inputs expected on manufacturing scale. The extent of protein polymerization was predicted to be within specifications for all simulated runs, demonstrating the robustness of the unit operation for subsequent process validation and future commercial manufacturing.

We illustrate how a Six Sigma project team can apply recursive partitioning to a historical data set to narrow down a list of potential experimental factors and then construct an experimental design using information from the partition analysis. The paper illustrates the value of analyzing historical manufacturing data to inform the choice of factors and levels for statistically designed experiments.

Methyl tert-butyl ether (MTBE) is a common groundwater contaminant, introduced to the environment by leaking petroleum storage tanks, urban runoff, and motorized watercraft. In this study. a simplified (static) headspace analysis method was adapted for determination of MTBE in water samples and soil water extracts. The MDL of the headspace method was calculated to be 2.0 microg L(-1) by the EPA single-concentration design method(1) and 1.2 microg L(-1) by a calibration method developed by Hubaux and Vos (Hubaux, A.; Vos, G. Anal. Chem. 1970,42, 849-855). The MDL calculated with the Hubaux and Vos method was favored because it considers both a true positive and a false positive. The static headspace method was applied to analysis of a tap water sample and a monitoring well sample from a gasoline service station, a river sample, and aqueous extracts from soil excavated during removal of a leaking underground storage tank (LUST). The water samples examined in this study had MTTBE concentrations ranging from 6 to 19 microg L(-1). Aqueous extracts of a soil sample taken from the LUST site had 8 microg L(-1) MTBE.

No abstract