Machine learning-driven screening of characteristic flavor markers: Investigating the impact of pulsed electric fields on the flavor profile of virgin olive oil.

Pulsed electric field (PEF) technology enhances the extraction efficiency of virgin olive oil (VOO), yet its systematic impact on the dynamic evolution of aroma profiles throughout processing remains unclear. This study investigated the effect of PEF on VOO quality across five critical processing stages and employed a machine learning (ML)-driven flavoromics approach to identify characteristic flavor markers. PEF significantly improved total phenol retention (up to 40.1%) and reduced oxidation indices. Key volatile markers revealed that PEF selectively increased the accumulation of lipoxygenase (LOX)-derived volatile compounds, increasing the odor activity value of the key green-note compound (E)-2-hexenal by 48.4%, while suppressing oxidation markers such as (E)-2-decenal to undetectable levels. A consensus ML framework (RF, SVM) achieved robust discrimination between PEF and traditional oils (AUC > 0.90), establishing (E)-2-hexenal and hexanol as reliable volatile fingerprints for process discrimination. These findings demonstrate that integrating flavoromics with ML provides a powerful strategy for optimizing PEF-assisted VOO production towards superior aroma quality.