Standardization of Raman spectra for transfer of spectral libraries across different instruments

Abstract

In this paper we evaluate methods for standardization of Raman spectra that are required to improve spectral correlation computations between spectra measured on different instruments. Five commercially-available 785 nm Raman spectrometers from different vendors were included in the study. These spectrometers have diverse specifications and performance levels and range in size from laboratory-based instruments to field-deployable portable and handheld platforms. Since each Raman spectrometer has different characteristics, spectra obtained on one instrument cannot readily be compared to a library acquired on a different instrument without performing various types of spectral corrections (standardization). We outline a procedure that combines previously established Raman shift and intensity correction protocols with a resolution matching step to facilitate the comparison of a centralized master library with spectra acquired on different geographically distributed Raman spectrometers. The standardization procedure is effective in reducing the inherent instrument-to-instrument variability so that spectra from different spectrometers can be compared and reliable results obtained using library-based spectral correlation methods. The findings have important implications for the ability to transfer Raman spectral libraries between instruments.

Standardization of milk mid-infrared spectra from a European dairy network

“…This is a drawback when recalibrating an instrument or using a historical database. Therefore, spectral corrections adapted to each instrument (standardization procedures) are needed (Rodriguez et al, 2011). One of the most common techniques for instrument standardization is the piecewise direct standardization (PDS) proposed by Wang et al (1991).…”

How the Critical Path Initiative Addresses CDER’s Regulatory Science Needs: Some Illustrative Examples

“…Use of these standardized spectra has allowed for much more reliable identification of drug products in the field. 5 Portable Raman spectrometers also have the potential to greatly improve the capacity of drug inspectors to identify products to which harmful adulterants have been added intentionally or inadvertently during manufacture. Because these adulterants may contribute relatively subtle changes to complex spectra, FDA researchers are developing and evaluating a variety of analytic procedures based on principal component analysis and soft independent modeling of class analogy to maximize the sensitivity and specificity with which the portable devices can detect adulterants such as diethylene glycol ( Figure 1).…”

“…The 7 topics from the CDER report are as follows: (1) improve access to postmarket data sources and their use in different types of analyses; (2) assessment and management strategies to reinforce the safe use of drugs; (3) evaluate the effectiveness and impact of different types of regulatory communications to the public and other stakeholders; (4) evaluate the links among product quality attributes, manufacturing processes, and product performance; (5) develop and improve predictive models of safety and efficacy in humans; (6) improve clinical trial design, analysis, and conduct; and (7) enhance individualization of patient treatment. ii.…”

Towards an improvement of model transferability for Raman spectroscopy in biological applications

“…With each chromosome, we receive a new wavenumber axis by fitting the values of the genes, i.e., the wavenumber adjustments of the control points, onto the whole wavenumber axis similarly as the standard wavenumber calibration. Afterwards, the hit-quality-index (HQI) [36] between the reference and the new Raman spectrum is calculated according to Eq.(1). This HQI is used as the evaluation function, which should be maximized during the algorithm evolvement.…”