Artificial Intelligence smelling based on sensomics: an effective tool to accurately quantify and monitor the most odor-active compounds in premium hazelnut (Corylus avellana L.) biscuits

Since its introduction, comprehensive two-dimensional gas chromatography (GC×GC), has unrevealed its potential in many fields helping scientists to better understand Nature’s complexity, facilitating highly-informative screenings, supporting markers discovery in omics applications, and offering many opportunities to implement system biology-like strategies for investigation [1]. The contribution deals with the challenging task of screening, identifying, and accurately quantifying character odorants in premium quality hazelnut biscuits in the context of industrial development of new products and their monitoring over time. In particular, as the Sensomics-Based-Expert-System (SEBES) introduced by Nicolotti et al. [2] acts as an Artificial Intelligence smelling machine, in the current study the GC×GC technology is combined with exhaustive multiple headspace extraction (MHE) and parallel detection by mass spectrometry and flame ionization detector (MS/FID) to enable prompt screening and accurate quantification of (key)-odorants in hazelnut biscuits. Key-odorants, validated by sensomics [3,4], and new potent odorants screened by GC olfactometry (GC-O) conducted by Aroma Extract Dilution Analysis (AEDA) on freshly prepared premium biscuits, are simultaneously analyzed and quantified with suitable accuracy. Moreover, by examining the matrix release effect on odorants, by MHE parameters, the textural and aroma retention properties of biscuits are evaluated supporting the technological development and product design.