Cell culture media and Raman spectra preprocessing procedures impact glucose chemometrics.

Area of Science:

• Process Analytical Technology (PAT)
• Chemometrics
• Spectroscopy

Background:

• Process Analytical Technology (PAT) tools like Raman spectroscopy are vital for pharmaceutical process automation and advanced manufacturing.
• Spectroscopic data preprocessing is a critical step influencing the performance of multivariate calibration models.
• Various preprocessing techniques exist, each impacting model development and predictive accuracy differently.

Purpose of the Study:

• To investigate the impact of different spectroscopic data preprocessing procedures on the development and performance of chemometric models.
• To optimize models for predicting glucose concentration in bioreactors using Raman spectroscopy.
• To compare the effectiveness of baseline correction (BLC), Savitzky-Golay smoothing (SGS), Savitzky-Golay derivative (SGD), and orthogonal signal correction (OSC).

Main Methods:

• Raman spectroscopy data were generated using a Box-Behnken design of experiment (DOE) considering glucose, glutamine, glutamic acid, and antifoam concentrations.
• Data were collected under varying aeration conditions and across three distinct cell culture media, including cell density variations.
• Partial Least Squares (PLS) regression models were developed using four preprocessing methods (BLC, SGS, SGD, OSC) individually and in combination.

Main Results:

• Orthogonal Signal Correction (OSC) preprocessing resulted in superior performance metrics for glucose prediction models, utilizing only one principal component across all media.
• Models employing BLC, SGS, or SGD required two or more principal components to achieve comparable performance to OSC.
• The choice of preprocessing procedure significantly influenced the overall performance and efficiency of the chemometric calibration models.

Conclusions:

• Preprocessing strategies critically affect the accuracy and efficiency of spectroscopic-based chemometric models in biopharmaceutical manufacturing.
• OSC emerges as a highly effective preprocessing technique for developing robust glucose concentration prediction models.
• Optimized preprocessing enhances process monitoring and control, contributing to advanced manufacturing capabilities.