The chemical composition of coffee beans varies according to the cultivar used, but there is little information about the differences among them resulting in different sensory profiles. The purpose was to find potential chemical markers to discriminate genealogical groups of coffee using regression analysis by partial least squares. Twenty-two accessions of C. arabica were used with two repetitions each. There were chromatographic and physical-chemical analyses to determine the composition of raw beans and chemometric analysis as the PCA and the PLS-DA. The analysis of PCA did not showed detailed information about the differences between the groups. The model PLS-DA identified the most important variables in the discrimination of the genealogical groups. Arachidic acid and stearic acid the markers for the Bourbon group; the myristic and linoleic acids and sucrose for the Exotic group; and lauric, palmitoleic and oleic acids, and the protein content to the Timor Hybrid group.
Keywords: Active germplasm bank; Arachidic acid PubChem CID: 10467); Caffeine (PubChem CID: 2519); Chemical markers; Coffea arabica L.; Discriminant analysis; Lauric acid PubChem CID: 3893); Linoleic acid (PubChem CID: 5280450); Linolenic acid (PubChem CID: 5280934); Myristic acid (PubChem CID: 11005); Oleic acid (PubChem CID: 445639); Palmitoleic acid (PubChem CID: 445638); Partial least square; Quality of coffee; Stearic acid (PubChem CID: 5281).; Sucrose (PubChem CID: 5988).
Copyright © 2020 Elsevier Ltd. All rights reserved.