Looking for Typical Traits in Monovarietal VOOs According to Their Phenolic Composition

Abstract

Due to its high sensitivity to numerous variability sources, it is hard to define the typicity of a monovarietal virgin olive oil (VOO) according to its phenolic profile. In this study, we aimed to identify the features of phenolic composition that are persistent and minimally affected by variability sources, making them potential varietal markers. We separately analyzed three databases of monovarietal VOO phenolic compositions, determined by liquid chromatography, from three different cultivars. The first database was produced from the original data of the Bosana cultivar. The other two were obtained through a systematic analysis of scientific literature on Coratina and Frantoio cultivars. Several statistical tools, including coefficient of variability, correlations, and linear regression models, were used to find recurring proportions or ratios unaffected by variability sources suitable to define typical varietal traits. Some proportions between molecules, mostly within the same phenolic class, remain constant. Strong correlations between (i) flavonoids were observed in Bosana and Frantoio VOOs (R² = 0.87 and 0.77, respectively), (ii) oleacein-oleocanthal (Bosana, R² = 0.81) (iii) oleuropein aglycon-ligstroside aglycon (Frantoio, R² = 0.88), and (iv) lignans (Coratina, R² = 0.84). These traits could be useful tools for defining the typicity of monovarietal VOOs.

Keywords: Coratina; Frantoio; Olea europaea L.; Pearson correlation; oleacein; oleocanthal; varietal characterization.

Figure 1

Main correlations and regression equations identified within the phenolic profile (mg kg⁻¹) of Bosana VOOs: p-coumaric acid–vanillic acid (a), oleocanthal–oleacein (b), ligastroside aglycon–oleuropein aglycon (c), apigenin–luteolin (d), tyrosol derivatives (sum of tyrosol, oleocanthal, and ligstroside aglycon)–hydroxytyrosol derivatives (hydroxytyrosol, oleacein, and oleuropein aglycon) (e).

Figure 2

Correlation matrix between the data acquired for 10 phenolic molecules, respectively, regarding Coratina VOOs (a) and Frantoio VOOs (b). Only the R values of significative Pearson correlations (p-value < 0.01) are reported in the figures.

Figure 3

Scatter plot reporting the linear regression model equations between individual phenolic molecules, potentially useful as varietal markers, for Coratina (a–d) and Frantoio (e–h) VOOs.

Figure 4

Summary plots of OPLS-DA models aimed at discriminating the Bosana VOOs according to different sources of variability: growing year (a–c), area of origin (d–f), and storage time (g–i). Figures (a,d,g) represent the observation scores colored by respective classes. Figures (b,e,h) represent the model loadings, showing the relationships among the X variables of the model. Figures (c,f,i) represent the variable importance on projections (VIPs) of the model, describing the importance of the variable to explain X and correlate with Y. Variable’s abbreviations: AcPin = 1-acetoxypinoresinol; Api = apigenin; Flav = flavonoids; HC = health claim; Hyd = hydroxytyrosol; Hyd_der = sum of Hyd, Oleac, Ol_Agl; LigAgl = ligstroside aglycon; Lut = luteolin; Ol_Agl = oleuropein aglycon; Oleac = oleacein; Oleo = oleocanthal; p-coum_ac = p-coumaric acid; Phen_ac = phenolic alcohols; Phen_alc = phenolic alcohols; Ratio_AC = Van_ac/p-coum_ac; Ratio_Agl = Ol_Agl/LigAgl; Ratio_Flav = Lut/Api; Ratio_O.O = Oleac/Oleoc; Sec = secoiridoids; TOT = sum of phenols; Tyr = tyrosol; Tyr_der = sum of Tyr, Oleoc, Lig_agl.; and Van_ac = vanillic acid.