April 4, 2019, Orlando, Florida. This year’s ACS National Meeting brought together industry and science experts in Orlando, just an hour west of the famous Space Coast. In the year of the 50th anniversary of the moon landing, one might think that curiosity, ingenuity and inquiring minds have also advanced other areas of science and technology. And indeed, they have.
Till Briskot and Tobias Hahn of GoSilico showed recent advances in chromatography modeling in two oral presentations and two posters. In the past months, GoSilico has focused on the development of a colloidal model for ion exchange chromatography and the application of Bayesian statistics for model calibration and scale-down model qualification.
The colloidal model we deduced can be used to simulate a wide pH range and has superior extrapolation power compared to previously used approaches. Based on molecular information, the behavior of the protein in the process is described by only three unknown parameters that can be identified by inverse modeling or even by QSAR approaches. We are confident that in the future only one experiment will suffice for model calibration, or perhaps not at all.
Insights on modeling and the permides project
The results obtained using Bayesian statistics were presented in the two oral presentations. We have developed a parameter estimation algorithm that uses a probabilistic surrogate function for the curve fitting objective. This reduces the number of iterations significantly to find the best parameter set and provides a confidence estimate. Together with prior knowledge about uncertain system properties, such as raw material variability, we use this information for risk-based scale-up and scale-down. In his talk, Till showed how our model-based predictions can be used to proof that scale-down models are representative for an at-scale process.
Finally, the results from PERMIDES project that was conducted in cooperation with CEVEC were presented in a poster. Here, we needed only two experiments to propose process conditions that achieved base line separation at a much higher load challenge.
The title pic was taken by Tobias at Kennedy Space Center.