“…The complexity of NMR data in food metabolomics studies is clearly the primary impetus for the coupling of NMR spectroscopy with multivariate statistical methods, capable of gathering samples with similar features Tomassini, Capuani, Delfini, & Miccheli, 2013). Among them, supervised methods that may enhance classification performance, such as orthogonal projection to latent structures -discriminant analysis (OPLS-DA) and its bidirectional modifications (O2PLS-DA) (Bylesjö, Rantalainen, Cloarec, Nicholson, Holmes, & Trygg, 2006), have shown great potential to determine the authenticity of various foodstuffs, mainly on the basis of their geographical or botanical origin (Consonni, Cagliani, & Cogliati, 2012a, 2012bConsonni, Cagliani, Stocchero, & Porretta, 2009Fotakis et al, 2013).…”