Evaluation of instrumental methods for the untargeted analysis of chemical stimuli of orange juice flavour

This work was conducted as the initial part of the evaluation of flavoromics as a tool in flavour research. The objective was to develop and evaluate methods for the untargeted analysis of chemical stimuli of orange juice flavour. It considered for study all (ideally) low molecular weight compounds as candidate chemical stimuli in flavour perception (unbiased) instead of focusing only on compounds already known to influence the flavour quality. Four commercial juices and their blend were analysed by headspace solid-phase micro-extraction gas chromatography (GC) and ultra-high-performance liquid chromatography (UHPLC)-time of flight mass spectrometry for volatiles and non-volatiles, respectively. The developed methods were a compromise between the number of compounds extracted and detected, throughput and repeatability. The methods were tested for their ability to distinguish between orange juices based on mass spectral information using chemometrics. Classification of the samples was not the goal of the study but rather an indirect way to test the instrumental methods, the handling and chemometric analysis of these data. Classification models were obtained which allowed the categorization of the samples by brands with little overlapping, and the tight clustering of the replicates indicated a good repeatability of the methods, especially for GC and RP-UHPLC. Fusion of GC- and RP-UHPLC-MS data sets gave similar classification models compared to that of using only data from volatiles or non-volatiles but can offer the advantage of finding potential correlations between chemical compounds and increased accuracy in flavour predictions as it includes inputs from more compounds.