Prediction of Chinese green tea ranking by metabolite profiling using ultra-performance liquid chromatography–quadrupole time-of-flight mass spectrometry (UPLC–Q-TOF/MS)

Metabolomics profiling provides comprehensive picture of the chemical composition in teas therefore may be used to assess tea quality objectively and reliably. In the present experiment, water and methanol extracts of green teas from China were analyzed by ultra-performance liquid chromatography–quadrupole time-of-flight mass spectrometry (UPLC–Q-TOF/MS) with the objectives to establish a model for quality prediction and to identify potential marker metabolites. The blindly evaluated sensory score of green teas was predicted with excellent power (R2=0.87 and Q2=0.82) and accuracy (RMSEP=1.36) by a partial least-squares (PLS) regression model based on water extract. By contrast, methanol extract failed to reasonably predict the sensory scores. The levels in water extract of neotheaflavin, neotheaflavin 3-O-gallate, trigalloyl-β-d-glucopyranose, myricetin 3,3′-digalactoside, catechin-(4α→8)-epigallocatechin and kaempferol were significantly larger whereas those of theogallin and gallocatechin were less in the low (score