Deciphering the Complexity of Smoke Point in Virgin Olive Oils to Develop Simple Predictive Models

The smoke point marks the onset of thermal degradation in edible oils. Although in this work we validated and improved its determination, it still relies on a subjective visual assessment and remains incompletely understood in relation to oil composition. This limitation reduces its reliability as a criterion for selecting frying oils in both industrial and culinary contexts. This study provides a systematic evaluation of how key chemical attributes of virgin olive oils influence their smoke point and proposes predictive models that could overcome the limitations of direct measurement. Forty-eight virgin olive oils were characterized, and multivariate modeling was applied to identify the most influential predictors. Free fatty acid content was the main determinant of the smoke point, exhibiting a strong inverse relationship, while saturated fatty acids and oxidative stability were shown to increase the smoke point by limiting the formation of volatile lipid oxidation products. Partial least squares models enabled accurate predictions using only routine quality parameters, such as free fatty acid content and saturated fatty acid content. Gaussian process regression further improved predictive performance and achieved high accuracy using free fatty acid content alone or, alternatively, other analytical parameters that are easily and routinely determined in olive oil. These findings offer a potential practical framework for estimating the smoke point without direct testing, with relevant implications for virgin olive oil quality control and the selection of oils for high-temperature applications.