Cannabis, an annual and dioecious member of the family Cannabaceae, is an economically important genus providing protein- and oil-rich seeds, long and short fibres for industrial applications (construction materials, textiles, paper, etc.), and a wide diversity of secondary metabolites found predominantly as terpenoids and cannabinoids (Lynch et al. 2016; McPartland 2018; Onofri and Mandolino 2017). In fact, the cannabis plant can produce over 150 unique terpenoids and roughly 100 unique cannabinoids, with a subset showing bona fide therapeutic utility (Hanuš et al. 2016; Booth and Bohlmann 2019). However, despite the large diversity in secondary metabolite profiles across thousands of cultivars, the stratification into drug-type cannabis or fibre-type cannabis hinges on the dry weight concentration of a single cannabinoid, Δ⁹-tetrahydrocannabinol (THC). This approach which prevails today in the USDA interim regulations, employs a THC concentration of 0.3% as the threshold separating hemp and drug-type cultivars, with concentrations below 0.3% defined as hemp (Dolgin 2019). Other jurisdictions outside North America have adopted higher thresholds, for example in Switzerland where concentration under 1% total THC is considered a compliant hemp crop. Sadly, despite human cultivation for over 6000 years in varying climates worldwide (Clarke and Merlin 2013), its evolution, taxonomic classification, and phylogenetic connections remain poorly understood. These deficiencies stem from limited research, irregular breeding efforts, unorganized selection, ex situ conservation, and government restrictions, which ultimately resulted in the high heterozygosity observed within cannabis genomes today (e.g. Rahn et al. 2016; McPartland 2018).

Although a subject of ongoing debate, taxonomic classification of the Cannabis genus has been delineated through three main types: 1) sativa (tall and less branched plant with long and narrow leaves), 2) indica (short and highly branched plant with broader leaves) and 3) ruderalis (heirloom with short stature, less branching and small thick leaves). While still under debate, particularly whether the genus is polytypic or monotypic, this broad classification reflects the putative geographical origins of each group (Clarke and Merlin 2013; Lynch et al. 2016, Schwabe and McGlaughlin 2019). Consequently, there is currently no structured horticultural registration system available for cannabis cultivars (varieties), instead these are often awarded the epithet “strains”, which are likely the outcome of extensive hybridization and subsequent rehybridization from their original botanical descriptors (Henry 2015).

The recent legalization of drug type cannabis for commercial production and recreational use in Canada, several US States, and other countries worldwide, as well a hemp under the US 2018 Farm Bill has created renewed scientific interest in developing a robust empirical classification system for cannabis. To that end, a particular focus has been placed on secondary metabolite expression with a clear separation based on CBD (cannabidiol):THC ratios. Differences in CBD:THC ratios delineate three class types: type-I (ratio < 0.5), type-II (ratio 0.5–3.0) and type-III (ratio > 3.0) (Elzinga et al. 2015), Interestingly, a genetic basis for these types can be determined by polymorphism at the CBDAS and THCAS genes on Chromosome 9 (Laverty et al. 2019). In addition, double recessives at this locus give rise to type-IV (cannabigerolic acid, CBGA accumulators; de Meijer and Hammond 2005) whereas type-V plants are free of cannabinoids which may have resulted from functionally ablative mutations founds within the upstream components of the cannabinoid synthase pathway (de Meijer et al. 2009). More recently, the addition of terpenoids as potential chemotaxonomic markers have emerged as a preferred model compared to cannabinoids alone (e.g. Lewis et al. 2018). Linking chemotype to genotype has also enabled deeper insight into a novel consumer-centric classification based on genetic markers associated with chemical expression (e.g. Orser and Henry 2019). Recently, others have proposed targeted markers for the identification of fiber and resin cannabis (e.g. Cascini et al. 2019; Hilyard et al. 2019) as well as molecular sexing tools to differentiate feminized from regular seed stock (Toth et al. 2020).

In addition to paving the way for empirical taxonomic classification, genetic information can provide insight into the extent and distribution of genetic variability, population structure, phylogenetic relationships, as well as providing the essential tools required to perform marker assisted selection in order to improved homozygosity and trait stability. In addition, genetic information can faithfully identify matching multilocus genotypes across disparate accessions. This feature may be particularly useful in seed-to-sale tracking as it provides an irrefutable identity for each individual accession and paves the way for cannabis variety registration and protection.

Here we describe a thorough investigation of cannabis accessions using a set of 23 highly informative and polymorphic SNP markers associated with important traits such as cannabinoid and terpenoid expression (Henry 2017; Henry et al. 2018; Orser and Henry 2019). We extend the scope of sampling to 420 accessions from Licenced Cultivators in Saskatchewan, Manitoba, British Columbia, Canada as well as Nevada, USA. We validate the use of these 23 SNP markers to assess population structure, phylogenetic relationship, population genetics, and correlation to secondary metabolite concentrations, and demonstrate the utility of this assay for rapid, repeatable and cost efficient genotyping of commercial and industrial cannabis accessions for use in product traceability, breeding programs, and consumer education.