Synthetic spectral libraries for Raman model calibration

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Biotechnology

Background:

• Raman spectroscopy is valuable for Process Analytical Technology (PAT) in bioprocessing.
• Model building for Raman spectroscopy is currently time-consuming and expensive.

Purpose of the Study:

• To explore the use of in silico spiking for creating synthetic spectral libraries (SSL).
• To reduce the cost and time associated with Raman spectroscopic model building.

Main Methods:

• Developed a fast calibration workflow using physical pure compounds spiked in water.
• Utilized in silico spiking of pure spectral fingerprints of various analytes.
• Created an SSL by fusing mammalian cell culture spectra, matrix variability, and pure component spectra.

Main Results:

• In silico spiking generated spectral data comparable to physical spiking.
• The SSL approach effectively reduced model building costs and time.
• Generated an extensive library of information-rich spectra.

Conclusions:

• In silico spiking is a viable and efficient method for enhancing Raman calibration.
• This approach provides a robust foundation for regression algorithms in bioprocess monitoring.
• The synthetic spectral library method significantly improves Raman spectroscopy model development.