Pre-processing of ultraviolet resonance Raman spectra

Abstract

The application of UV excitation sources coupled with resonance Raman have the potential to offer information unavailable with the current inventory of commonly used structural techniques including X-ray, NMR and IR analysis. However, for ultraviolet resonance Raman (UVRR) spectroscopy to become a mainstream method for the determination of protein secondary structure content and monitoring protein dynamics, the application of multivariate data analysis methodologies must be made routine. Typically, the application of higher order data analysis methods requires robust pre-processing methods in order to standardize the data arrays. The application of such methods can be problematic in UVRR datasets due to spectral shifts arising from day-to-day fluctuations in the instrument response. Additionally, the non-linear increases in spectral resolution in wavenumbers (increasing spectral data points for the same spectral region) that results from increasing excitation wavelengths can make the alignment of multi-excitation datasets problematic. Last, a uniform and standardized methodology for the subtraction of the water band has also been a systematic issue for multivariate data analysis as the water band overlaps the amide I mode. Here we present a two-pronged preprocessing approach using correlation optimized warping (COW) to alleviate spectra-to-spectra and day-to-day alignment errors coupled with a method whereby the relative intensity of the water band is determined through a least-squares determination of the signal intensity between 1750 and 1900 cm(-1) to make complex multi-excitation datasets more homogeneous and usable with multivariate analysis methods.

UV Resonance Raman Investigations of Peptide and Protein Structure and Dynamics

“…126 They also used constrained multivariate curve resolution-alternating least squares (MCR-ALS) analysis 127 and parallel factor (PARAFAC) analysis 126 to resolve discrete protein secondary structural motifs.…”

Spectral pre-processing for biomedical vibrational spectroscopy and microspectroscopic imaging

“…Most of the modern dispersive Raman instruments utilize gratings in combination with sensitive CCD detectors for collecting the weak Raman signals. One of the main problems of such multichannel Raman spectrometers is the problem of wavelength stability [77,78]. Wavelength inaccuracies are often times inevitable and may occur as a result of instrumental factors such as source/grating changes, misalignment of the collection optics, thermal changes, and other factors.…”

“Parallel factor analysis of multi-excitation ultraviolet resonance Raman spectra for protein secondary structure determination”

“…Spectra collected at different excitation wavelengths have different resolutions. [44] Each spectrum was interpolated to match the number of variables in the 197 nm spectra. Phosphate buffer spectra were subtracted from each protein spectra.…”

“…[39] Initially, quantification studies of protein secondary structure employing DUVRR focused on univariate calibration methods of single amide modes, [26] but have evolved to include multivariate calibration and multivariate curve resolution methods and other advanced statistical analyses of all observable amide modes. [22,35,36,[40][41][42][43][44][45] In these methods, first order data are decomposed using bilinear models showing that individual protein spectra, x, are a linear combination of the different underlying pure secondary structure motifs, s, and their respective fractional amounts, c (Equation 1);…”