5.2.2. Optimizing factors using response surface methodology
Based on PBD results, four of the most influential factors were selected
for further optimization by CCD in DoE software, resulting in thirty
experimental runs (Supplementary table 2). The specified composition of
model terms (chosen factors) was employed to prepare culture media for
each run according to the designed points via BioLector. Besides,
constant values of other media components not in the model were
supplemented in the media according to the insulin results (mg/ml) in
PBD design and SDS-PAGE results. The concentration corresponding to the
highest response for less significant variables and -1 level of
insignificant factors were supplied. Following the execution of
experiments via BioLector, responses were scrutinized via different
models. The best model was picked based on model validation parameters
documented in the ANOVA table, fit statistic tables, and diagnostic
analysis. DoE software generated the perturbation chart, and the effect
of each significant independent and dependent variable on response was
registered graphically via contour and 3D plots. Finally, DoE software
generated four scenarios about optimal points based on the acquired
regression equation. Predicted design scenarios with the highest
desirability were dissected and compared to each other using Biolector
XT microbioreactor in triplicates again. The suggested optimal media was
defined as Scenario-3 (13 mg/ml) with a 3.3 deviation and is ready for
scale-up optimization.
Acknowledgments
The authors dedicate this manuscript to the memory of Albert E. Dahlberg
and Nizar Turker. The authors gratefully acknowledge the use of the
services and facilities of the University of Health Science-Validebag
DETAUM (Experimental Medicine Research and Application Center).