2D NMR spectroscopy in food science: an innovative approach to spectra processing and statistical data analysis

NMR spectroscopy is nowadays applied to a wide range of matrices, regardless of their physical state, with some relevant advantages: samples remain unchanged, it is possible to obtain information regarding different class of molecules with one acquisition and the production of dangerous waste is avoided. The use of 1H NMR in foods is of particular interest, especially when coupled with chemometrics and multivariate statistical analysis, techniques that have been applied to quality control studies and in the assessment of botanical and geographical origin of foods. This kind of procedure has been applied to many products, from oils to honey (Bertelli et al. 2010), from beverages like beer, tea and coffee to oenological products like wine and vinegar (Bertelli et al 2012). Nevertheless, the majority of those applications regards monodimensional NMR, since this kind of spectra is usually faster to acquire and easier to use, while 2D spectra often require more complex procedures. To overcome this limitation, we are developing a tool for the analysis of bidimensional spectra, able to import and analyse them using Matlab environment. This software is able to use instrumental data as raw signals, by unfolding each spectra into a vector, as well as a whole image, using a pattern recognition procedure. To evaluate the applicability and the effectiveness of this approach, a set of PDO Lambrusco wines and of Balsamic and Traditional Balsamic Vinegar of Modena have been analysed, acquiring 1H-13C HMBC spectra using a 400 NMR spectrometer. The obtained spectra were first processed by means of the developed software, then generated datasets have been analysed using multivariate statistical methods. Results shows that the tool, even if still under development, is able to exploit 2D NMR spectra, making possible to build classificatory models using extracted informations.