An untargeted metabolomic approach in the chemotaxonomic assessment of two Salvia species as a potential source of α-bisabolol.

α-Bisabolol is a commercially important aroma chemical currently obtained from the Candeia tree (Vanillosmopsis erythropappa). Continuous unsustainable harvesting of the Candeia tree has prompted the urgent need to identify alternative crops as a source of this commercially important sesquiterpene alcohol. A chemotaxonomic assessment of two Salvia species indigenous to South Africa is presented and recommended as a potential source of α-bisabolol. The essential oil obtained by hydrodistillation of the aerial parts was analysed by gas chromatography coupled to mass spectrometry (GC-MS) and mid-infrared spectroscopy (MIRS). Orthogonal projections to latent structures-discriminant analysis (OPLS-DA) were used for multivariate classification of the oils based on GC-MS and MIRS data. Partial least squares (PLS) calibration models were developed on the MIRS data for the quantification of α-bisabolol using GC-MS as the reference method. A clear distinction between Salvia stenophylla and Salvia runcinata oils was observed using OPLS-DA on both GC-MS and MIRS data. The MIR calibration model showed high coefficient of determination (R² = 0.999) and low error of prediction (RMSEP=0.540%) for α-bisabolol content.