Liquid chromatographic fingerprints and profiles of polyphenolic compounds applied to the chemometric characterization and classification of beers

In this paper, liquid chromatography with UV-vis detection was used to generate compositional fingerprints of beers to be exploited for characterization and classification purposes. Chromatographic profiles recorded at 280 nm contained features mainly associated with polyphenolic components such as phenolic acids and flavonoids. Beers of different styles and brewed in various countries were analyzed by the proposed method and the data generated were treated chemometrically to assess characterization and classification models. Three different types of data sets based on chromatograms, peak areas and concentrations were explored by principal component analysis (PCA) to evaluate their performances to discriminate among ale and lager beers. The use of raw chromatographic profiles required a comprehensive pretreatment to improve the data quality. When dealing with peak areas, single and complex integrated peaks of known and/or unknown compounds were used as the source of analytical information. In these two approaches (chromatographic fingerprints and peak areas), calibration was not necessary so the sample analysis was simplified. In the case of concentrations, selected phenolic acids and flavonoids were considered as the data to discriminate among beer types. Differences in the polyphenolic composition were relevant and some components resulted in efficient markers of beer classes. Further studies based on partial least squares discriminant analysis (PLS-DA), soft independent modelling of class analogy (SIMCA) and other methods were used to discriminate beers according to brewing styles. Classifications were highly satisfactory in terms of selectivity and sensitivity as, in general, beers of test sets were correctly assigned to their actual classes.