Molecular sensory science combined with machine learning for exploring key odor-active compounds and aging-feature compounds of Shanxi aged vinegar.

Aroma compounds in fermented foods are from raw material or produced by microorganisms, contributing to their diverse aroma profiles. The key odor-active compounds (OACs) present in Shanxi aged vinegar (SAV), one of China's famous vinegars, were systematically analyzed by molecular sensory science. A total of 152 OACs were identified in this study. Moreover, 87 OACs were quantitatively determined, and 37 OACs exhibited their odor activity values (OAVs) larger than 1. The aroma recombination and omission experiments confirmed that ketones, pyrazines, lactones, and acids are the most significant contributors to the aroma profiles of SAV. The correlation analysis confirmed that the absence of specific aroma compounds not only directly impacts the sensory characteristics associated with the missing compounds but also influences other sensory attributes. Therefore, 47 OACs, classified as key OACs through the verification of recombination and omission test, based on AEDA (FD ≥ 400) or OAV (≥1). Lastly, 6 aging-related markers (methional, 1-dodecanol, acetoin, benzyl acetate, propanoic acid, and trimethyl-pyrazine) were determined by machine learning based on key OACs. These results provide a standard for screening the key OACs that determine aroma characteristics of SAV, as well as for product quality control.