The two first components accounted for 94.2% of the total sample variance.

A clear separation between defective and non-defective coffees can be observed, with four major groups: non-defective (positive PC1, negative PC2), GW3965 immature/light sour (positive PC1, positive PC2), dark sour (negative PC1, positive PC2), and black (negative PC1, negative PC2). Clustering of immature and light sour defects was also observed by Mendonça et al. (2008) for analysis of the ESI(+)-MS profiles. Evaluation of the loadings plot (not shown) indicated that the spectral ranges that presented the highest influence on sample grouping were 2980–2850 and 1560–800 cm−1 corresponding to immature/light sour samples, 1700–1570 cm−1 corresponding to non-defective beans, 3100–3000, 1980–1760 in reference to dark sour, and 2000–1985 cm−1 corresponding to black beans. Group separation was not enhanced by taking derivatives of the spectra. PCA analysis of the ATR reflectance spectra, employing baseline correction and normalisation is displayed in Fig. 4b. Analysis was based on a 54 × 1188 data matrix assembled so that each row corresponded to a sample and each column represented the spectra data at a given wavelength. The two first components accounted for 86.3% of the total sample variance. The

first component provided separation of the evaluated samples into two major groups: non-defective/light sour (positive PC1) and black/dark sour/immature (negative PC1). Evaluation of the loadings plot (not shown) indicated that the spectral ranges that presented the highest influence on http://www.selleckchem.com/products/tenofovir-alafenamide-gs-7340.html PC1 values in association with the black/dark sour/immature group were the following: 1554–1482, 1797–1776 and 3100–3020 cm−1. The only significant band that can be observed in the ATR spectra (Fig. 2b) in those ranges

is the one at 1534 cm−1. It was also present in the spectra obtained by Lyman et al. (2003) for aqueous extracts of roasted coffee, regardless of roasting conditions, although no identification was attempted. We herein infer that this band is most likely associated to C C stretch in a nitrogen based ring, such as caffeine or trigonelline, both mafosfamide present in significant amounts in raw and roasted coffees. In the case of PCA based on the first-derivative of the spectra, the first and second principal components accounted for 28.6% and 12.6% of the total sample variance, respectively. No group separation could be observed. To verify the similarity between the samples and to single out some classes, agglomerative hierarchical clustering (AHC) was applied to the set of variables employed for PCA. Ward’s hierarchical clustering method and chord distance were employed to establish clusters and calculate dissimilarity coefficients, respectively. The resulting dendrograms are displayed in Fig. 5. Group clustering confirmed the PC analysis. In the case of transmittance spectra obtained with KBr discs (Fig.