Standardising the chromatographic denoising procedure. Issue 34 (21st August 2018)

• Record Type: Journal Article
• Authors: Kumar, Keshav
• Is Part Of: Analytical methods. Volume 10:Issue 34(2018)
• Page Start: 4189
• Page End: 4200
• Publication Date: 2018-08-21
• Subjects: Chemistry, Analytic -- Periodicals; Analytical biochemistry -- Periodicals; Chemical laboratories -- Standards -- Periodicals; 543.1905
• DOI: 10.1039/c8ay01606b
• Languages: English
• ISSNs: 1759-9660

Abstract

Abstract : The present work standardizes the chromatographic denoising procedure by comparing Wavelet analysis, chemometrics techniques and Savitzky–Golay (SG) algorithm. Abstract : The denoising of chromatograms is an important step before subjecting them to any data analysis workflow. The present work demonstrates a comparison between the wavelet analysis, Savitzky–Golay (SG) algorithm and chemometric approaches towards finding a simple, fast and user-friendly denoising approach. It is found that the wavelet analysis and SG algorithm require the optimisation of several parameters making the entire denoising procedure user involved, laborious and time consuming. Moreover, the denoising of the two chromatographic data sets achieved with the wavelet analysis and SG algorithm is not at the desired level. Compared to the wavelet analysis and SG algorithm, chemometric techniques are found to provide a better denoising of chromatograms. In the present work, two chemometric techniques principal component analysis (PCA) and multivariate curve resolution alternating least squares (MCR-ALS) analysis are tested. Between these two chemometric techniques, PCA is found to provide a simple and better denoising approach. The application of PCA does not require the optimisation of any user specified parameters. PCA also provides better denoising, which is mainly because it involves the projection of the data sets in the space spanned by principal components that are essentially orthogonal to each other and most of the important information is mainly captured by a few initial principal components. The noise contents of the data sets are usually explained by higher principal components. Thus, the application of PCA with a suitable number of components denoises the chromatograms while preserving all the essential information. The present work, clearly suggests that PCA can provide an easy, fast and simple denoising procedure and hence must be integrated in the data analysis workflow.