Characterisation of Sclerocarya birrea (marula) seed oil and investigation of the geographical origin by applying similarity calculations, differential NMR and hierarchical cluster analysis

IntroductionThe marula fruit is an important indigenous African fruit since various commercial products are produced from the pulp and the seed oil. The increased demand requires methods for authentication, quality control and determination of geographical origin. ObjectiveThe study aimed to establish a fast and reliable method for characterisation and authentication of marula seed oil. Furthermore, to identify marker compounds that can distinguish marula seed oils from other commercial oils and indicate regional differences. Materials and MethodsMetabolic profiling of 44 commercial marula seed oils was performed using proton nuclear magnetic resonance (H-1 NMR). For rapid classification similarity calculations were compared with principal component analysis. Differential NMR was used to determine marker compounds. ResultsMarula seed oil was found to be similar to macadamia and olive oils and was distinguished from these oils by the detection of minor components. Marula seed oil is differentiated from the other two oils by the absence of & alpha;-linolenic acid, relatively high levels of monoglycerides and diglycerides, and an approximately 1:1 ratio of 1,2- and 1,3-diglycerides. When comparing marula seed oils from various regions using hierarchical cluster analysis, clustering of the marula seed oils from Namibia and Zimbabwe was observed and was related to the quantities of linoleic acid and monoglycerides and diglycerides. Some samples displayed deviations in their composition which might indicate adulteration or contamination during the production process. ConclusionThe study demonstrates the potential of NMR as a tool in the quality control of marula seed oil. This technique requires very little sample preparation, circumvents derivatisation of the oil components with fast run-times. In addition, samples with chemical profiles that differ from the general signature profile can easily be identified.