Exploration of liquid chromatographic-diode array data for Argentinean wines by extended multivariate curve resolution

Abstract

22Second-order data were measured using high-performance liquid-chromatography 23 with diode array detection (HPLC-DAD) for a number of wine samples, which were directly 24 injected in the HPLC-DAD system without sample pre-treatment. The data were arranged in 25 data matrices whose modes were elution time and UV-visible absorption wavelength, and 26 processed by extended multivariate curve resolution coupled to alternating least-squares 27 (MCR-ALS). The individual data matrices were organized in a row-wise augmented data 28 matrix sharing the time subspace, due to the high spectral similarity among several sample

Anthocyanins as markers for the classification of Argentinean wines according to botanical and geographical origin. Chemometric modeling of liquid chromatography–mass spectrometry data

“…As mentioned above, in most reported works using elution profiles as fingerprint for wine classification, it is very difficult to identify key compounds within that profile that contributed most to achieve the classification when no pre-treatment was applied to the wine sample. In concordance with our previous work (Pisano et al, 2014), in which we conclude that different anthocyanin compounds contributed to wine discrimination, in this work we found that anthocyanin compounds that contribute to discrimination were three malvidin-derived anthocyanins, considering the remarkable fact that the study was performed by direct injection of untreated wine samples.…”

“…Previously in our working group we performed an exploration of Argentinean red wines by direct injection HPLC-DAD without sample pre-treatment coupled to multivariate curve resolutionalternating least-squares (MCR-ALS) as a chemometric data processing algorithm (Pisano, Silva, & Olivieri, 2014). We also attempted wine classification by grape varietal and geographical origin, achieving Malbec varietal discrimination from the remaining ones, and partial success in discriminating samples according to their geographical origin.…”

Phenolic Compounds as Markers of Wine Quality and Authenticity

“…In addition, other researchers suggested combining the profile of anthocyanins with the content of phenolic acids [61], flavonols [18], phenolic acids, and flavan-3-ols [62]. Based on the statistical evaluation, flavan-3-ols alone [63] or combined with phenolic acids [64] and condensed proanthocyanidins [65,66] were found to distinguish well Graciano, Tempranillo, Cabernet Sauvignon, Cabernet Franc, Carménère, Merlot Pinotage, Syrah, and Sangiovese grape varieties.…”

Aligning retention time shifts in HPLC three‐dimensional spectra by icoshift approach combined with data arrangement methods and the release of a graphical user interface

“…Here, we focus our attention on the later one. In the column-wise augmentation, the data matrices are placed on top of each other (keeping the same number of columns in all of them), and in case of rowwise augmentation, the individual data matrices are placed one adjacent to other (keeping the same number of rows in all of them) [18]. For example, the multi-compounds in table wines were successfully quantified on the basis of 1 H NMR SC-3D spectra that were established by the repetition of the original corresponding 1 H NMR spectrum [14]; the amino acids in the yellow foxtail millet substrate were determined on the basis of the terahertz time-domain spectroscopy (THz-TDS) images that include the spectra of three different parameters [15]; three components in mixture samples were quantitatively analyzed on the basis of near-infrared (NIR) 3D spectra with temperature [16].…”

“…Moreover, the other ways for data arrangement have also been reported, such as multivariate curve resolution-alternating least squares (MCR-ALS) method can be easily extended to the simultaneous analysis of multiple data sets through column-wise or row-wise augmented data matrices [17]. In the column-wise augmentation, the data matrices are placed on top of each other (keeping the same number of columns in all of them), and in case of rowwise augmentation, the individual data matrices are placed one adjacent to other (keeping the same number of rows in all of them) [18]. One can conclude that the above data arrangement ways may be conductive to use the data flexibly to meet different analysis requirements, and other novel data arrangement methods are also worthy of expectation.…”