Clustering Monovarietal Extra Virgin Olive Oil According to Sensory Profile, Volatile Compounds, and k-Mean Algorithm

A real valorization of monovarietal extra virgin olive oils (EVOOs) through diverse sensory profiles is missing, hindering different uses of samples with different sensory profiles and leading to negative impacts on EVOO competitions and consumer choices. The aim of this research is to group monovarietal EVOOs according to similar sensory and chemical profiles. The volatile and phenolic composition and the sensory profile of 46 monovarietal EVOOs are analyzed by head space-solid phase micro extraction-gas chromatography-mass spectrometry, HPLC-DAD, and Panel Test; the data are used to feed a k-mean algorithm aimed at samples clustering. A group of non-monovarietal samples is also included. It is hypothesized that samples of a cultivar are located in a cluster, while non-monovarietal samples are randomly located in different clusters. Two clusters of samples are identified; all samples belonging to a specific cultivar are in the same cluster and the non-monovarietal ones randomly placed in the two clusters. The significant attributes in differentiating between the two clusters belong to sensory descriptors and volatile compounds; phenolic compounds do not give significant differentiation. All sensory descriptors result in prevalence in one cluster. The two clusters are differentiated by volatile organic compounds (VOCs) originating by different branches of the lipoxygenase-pathway. Practical Applications: The results of this research will help toward a real valorization of extra virgin olive oils (EVOOs), and particularly, the monovarietal ones. In this sense, the results clearly suggest the use of an adequate profile sheet for the definition of the sensory profile of monovarietal EVOOs. By this way, during EVOO competitions, samples with a certain type of sensory profile will compete against samples with the same type of sensory profile, just as it happens for other types of products.