Development of Raman Spectroscopy and Machine Learning Methods for Protein Aggregate Quantification: Application to...

Area of Science:

• Biopharmaceutical Manufacturing
• Analytical Chemistry
• Process Analytical Technology

Background:

• Protein aggregation is a critical issue affecting biopharmaceutical product quality, efficacy, and safety.
• Traditional offline methods for aggregate detection, like size-exclusion chromatography, lack the speed for real-time process control.
• There is a significant need for inline analytical techniques to monitor protein monomers and aggregates during manufacturing.

Purpose of the Study:

• To develop and validate a Raman spectroscopy-based strategy for the selective detection and quantification of protein monomers and aggregates.
• To address the limitations of traditional methods by enabling real-time monitoring in biopharmaceutical processes.
• To utilize advanced chemometric approaches for robust quantification of protein size variants.

Main Methods:

• Raman spectroscopy was employed for molecular specificity and rapid data acquisition.
• Controlled stress conditions were used to generate reproducible protein aggregates (bovine serum albumin).
• A Latin Hypercube sampling design varied protein concentration and aggregate fraction to isolate aggregation effects.
• Convolutional neural networks (CNNs) were utilized as a chemometric machine learning approach for data analysis.

Main Results:

• Spectral markers indicative of protein aggregation were identified.
• CNNs demonstrated superior predictive performance and robustness in quantifying monomers and aggregates compared to traditional methods.
• The developed method achieved reliable, real-time monitoring of protein size variants.
• Qualitative comparison with offline size-exclusion chromatography validated spectral marker findings.

Conclusions:

• Raman spectroscopy, coupled with advanced chemometric modeling (specifically CNNs), provides a reliable method for real-time monitoring of protein aggregation.
• This approach significantly enhances biopharmaceutical downstream process control and robustness.
• The findings support the application of Raman spectroscopy as a Process Analytical Technology (PAT) for quality control in biopharmaceutical manufacturing.