From Omics to Industrial Assessment: Validation of GC×GC-Derived Oxidation Markers for Hazelnut Shelf-Life Prediction

The European hazelnut (Corylus avellana L.) is a high-value raw material widely used in the confectionery industry, where product quality is traditionally assessed through visual inspection and sensory evaluation. These methods, while effective, are inherently subjective and limited in their ability to detect early-stage deterioration or provide predictive insights. A modern, science-driven concept of food quality should integrate molecular-resolution approaches capable of enriching conventional evaluations with objective, quantifiable data related to authenticity, stability, and technological properties. Among the most informative molecular signatures in hazelnuts is the volatilome, i.e. the volatile metabolome, which encodes a wealth of quality-related information. This includes markers of cultivar/geographical origin, post-harvest treatments, microbial contamination, oxidative degradation, and sensory characteristics. The peculiar distribution of potent odorants within this complex mixture defines the sensory identity of the product, making volatilomics a powerful tool in food quality profiling. In this study, a Sensomics-Based Expert System (SEBES), an AI-driven workflow for aroma decoding [1], was employed and enhanced through the use of comprehensive two-dimensional gas chromatography coupled with mass spectrometry and flame ionization detection (GC×GC-MS/FID), along with multiple headspace solid-phase microextraction (MHS-SPME). This integrated approach enables high-throughput and accurate quantification of about 40 volatiles, including key aroma compounds, spoilage indicators, and geographical tracers. Monitoring hazelnut samples across four harvest years, several volatile compounds showed strong and consistent correlations with lipid oxidation phenomena, a key parameter in predicting shelf life and ensuring industrial product quality. This contribution focuses on the validation of these potential oxidation markers, for which dedicated autoxidation experiments were carried out using model systems based on pure triacylglycerols (TAGs) and custom mixtures that simulate the lipid composition of raw hazelnuts. These systems were exposed to different temperatures and oxygenation levels to simulate realistic storage and processing conditions. The analytical results confirmed the formation of several well-known oxidation products such as (E)-2-Nonenal, Heptanal, (E)-2-Decenal, Nonanal, Octanal, 1-Octanol, and 1-Octen-3-ol as well as less commonly reported compounds including Octanoic acid, Heptanoic acid, 1-Hexanol, and 1-Pentanol. These findings support the use of selected volatiles as reliable early indicators of oxidative deterioration, offering a powerful, objective tool for predicting shelf life and enhancing quality control strategies in hazelnut processing. Beyond the specific case of hazelnuts, this study underscores the broader potential of volatilomics and high-separation-capacity analytical platforms, supported by Artificial Intelligence, to define the next generation of advanced, data-driven food quality evaluation systems.