Abstract: In order to investigate the contents and distribution patterns of saccharides and amino acids in five types of honey from Yunnan province, Macadamia integrifolia honey, Coffea honey, Leucosceptrum canum Smith honey, Vicia villosa Roth honey, and Hevea brasiliensis honey, the contents of 12 saccharides and 20 amino acids in them were determined, and hierarchical cluster analysis (HCA) and partial least squares-discriminant analysis (PLS-DA) were used to explore the applicability of saccharide and amino acid contents to their classification and discrimination. The results showed that the contents of fructose, glucose and sucrose in the five types of honey met the requirements of the industry standards. The total amount of 10 oligosaccharides ranged from 5.35% to 7.92%, with a significant correlation being observed between turanose and mlatulose, kojibiose and isomaltose. The total amount of 20 amino acids ranged from 312.54 to 1 245.51 mg/kg, in the decreasing order of Coffea honey > L. canum Smith honey > V. villosa Roth honey > H. brasiliensis honey > M. integrifolia honey. Significant differences in amino acid composition and content were found among honey types, and Coffea honey contained more than 10 times as much tyrosine as all other types, from which it was inferred that tyrosine could be a potential marker of Coffea honey. The results of cluster analysis showed good classification of the five types of honey according to amino acids with an accuracy of 98.5%. PLS-DA could also distinguish them effectively, and six differential amino acids, proline, arginine, tryptophan, histidine, asparagine and alanine, were selected using the cutoff of variable importance in projection (VIP) value > 1, which could be used to classify the five types of honey in Yunnan.

Key words: honey; sugar; amino acid; chemometrics