Introduction: With laws changing around the world regarding the legal status of Cannabis sativa (cannabis) it is important to develop objective classification systems that help explain the chemical variation found among various cultivars. Currently cannabis cultivars are named using obscure and inconsistent nomenclature. Terpenoids, responsible for the aroma of cannabis, are a useful group of compounds for distinguishing cannabis cultivars with similar cannabinoid content. Methods: In this study we analyzed terpenoid content of cannabis samples obtained from a single medical cannabis dispensary in California over the course of a year. Terpenoids were quantified by gas chromatography with flame ionization detection and peak identification was confirmed with gas chromatography mass spectrometry. Quantitative data from 16 major terpenoids were analyzed using hierarchical clustering analysis (HCA), principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), and orthogonal partial least squares discriminant analysis (OPLS-DA). Results: A total of 233 samples representing 30 cultivars were used to develop a classification scheme based on quantitative data, HCA, PCA, and OPLS-DA. Initially cultivars were divided into five major groups, which were subdivided into 13 classes based on differences in terpenoid profile. Different classification models were compared with PLS-DA and found to perform best when many representative samples of a particular class were included. Conclusion: A hierarchy of terpenoid chemotypes was observed in the data set. Some cultivars fit into distinct chemotypes, whereas others seemed to represent a continuum of chemotypes. This study has demonstrated an approach to classifying cannabis cultivars based on terpenoid profile.
Keywords: chemotype; hierarchical clustering analysis; orthogonal partial least squares discriminant analysis; partial least squares discriminant analysis; principal component analysis; terpenes.