Untargeted metabolomics with multivariate analysis to discriminate hazelnut (Corylus avellana L.) cultivars and their geographical origin.J Sci Food Agric. 2020 Jan 30; 100(2):500-508.JS
In the present study a metabolomics-based approach was used to discriminate among different hazelnut cultivars and to trace their geographical origins. Ultra-high-pressure liquid chromatography coupled to quadrupole-time-of-flight mass spectrometry (UHPLC-ESI/QTOF-MS) was used to profile phenolic and sterolic compounds.
Compounds were identified against an in-house database using accurate monoisotopic mass and isotopic patterns. The screening approach was designed to discern 15 hazelnut cultivars and to discriminate among the geographical origins of six cultivars from the four main growing regions (Chile, Georgia, Italy, and Turkey). This approach allowed more than 1000 polyphenols and sterols to be annotated. The metabolomics data were elaborated with both unsupervised (hierarchical clustering) and supervised (orthogonal projections to latent structures discriminant analysis, OPLS-DA) statistics. These multivariate statistical tools allowed hazelnut samples to be discriminated, considering both 'cultivar type' and 'geographical origin'. Flavonoids (anthocyanins, flavanols and flavonols - VIP scores 1.34-1.49), phenolic acids (mainly hydroxycinnamics - VIP scores 1.35-1.55) together with cholesterol, ergosterol, and stigmasterol derivatives (VIP scores 1.34-1.49) were the best markers to discriminate samples according to geographical origin.
This work illustrates the potential of untargeted profiling of phenolics and sterols based on UHPLC-ESI/QTOF mass spectrometry to discriminate hazelnut and support authenticity and origin. © 2019 Society of Chemical Industry.