The geographical origin of pulses influences their chemical composition and functional properties, due to environmental and agronomic factors. In this study, gas chromatography coupled with Q-Orbitrap-high-resolution mass spectrometry (GC-Q-Orbitrap-HRMS) was applied to explore compositional differences between Spanish and Southern Italian bean samples using a non-targeted metabolomic approach. Volatile compounds were profiled by headspace solid-phase microextraction (HS-SPME), while non-polar semi-volatile components were investigated using salt-assisted liquid-liquid extraction (SALLE) followed by direct injection. Multivariate statistical analyses, including principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA), revealed partial group structuring under unsupervised conditions and satisfactory supervised model performance within the investigated dataset (R2Y > 0.8; Q2 > 0.6). An exploratory mid-level data fusion strategy was implemented to integrate complementary analytical features from both datasets and examine its effect on model performance and variable ranking. Overall, seven metabolites (e.g., 1-hexadecanol, 3,5-octadien-2-one, hexadecane) were putatively annotated in the fused dataset. They contributed to the supervised differentiation of Southern Italian and Spanish beans, highlighting origin-related variability within the metabolomic fingerprint of geographically related legume matrices.
