Massively parallel sequencing (MPS), also called next generation sequencing (NGS), is a high throughput technique capable of collecting DNA sequence data from multiple targets and multiple samples in parallel. It offers several distinct advantages over traditional DNA typing using capillary electrophoresis (CE), including providing more comprehensive coverage of target markers (sequence data in addition to length) and the ability to analyze hundreds or thousands of targets at a time (Bruijns et al. 2018; Moorthie et al. 2011), compared to only about 25 loci by CE for a five-dye short tandem repeat (STR) kit (Lazaruk et al. 1998). Sequence data may elucidate isoalleles, alleles which have the same length and appear identical on CE but actually have different sequences, leading to more discriminatory results. Isoalleles may differ in their repeat structure or contain variants in the flanking regions, and the International Society of Forensic Genetics (ISFG) has reported guidelines to standardize the nomenclature for these sequence variable alleles (Parson et al. 2016). Costs and run times associated with MPS have dropped substantially, making targeted MPS assays a cost-effective approach for characterizing samples for genetic individualization or identification (Moorthie et al. 2011).

In forensics, MPS assays have been used for human identification purposes, including sequencing autosomal STRs and single nucleotide polymorphisms (SNPs) (Eduardoff et al. 2015; Guo et al. 2017; Kim et al. 2016; Seo et al. 2013; Wang et al. 2017), mitochondrial DNA analysis (Davis et al. 2015), phenotype prediction (Mehta et al. 2016), and other purposes (Bruijns et al. 2018; Budowle et al. 2017). While much of the forensic research on MPS has focused on human DNA, its use for forensic plant science has recently been investigated (Houston et al., 2018a, b). Houston et al. reported an MPS panel for the Ion S5™ consisting of twelve autosomal STRs in Cannabis sativa (marijuana). Results showed concordance with CE methods, and isoalleles were found at eight of the loci, providing a higher discriminatory power compared to CE (Houston et al., 2018a, b). MPS has been also been used for DNA barcoding studies in animals and has shown better recovery, reduced costs, and faster processing times compared to traditional Sanger sequencing (Shokralla et al. 2015). Studies involving the use of MPS for DNA barcoding in plants have been limited but do show the advantage of simultaneous analysis of multiple barcodes for enhanced phylogenetic resolution (Parks et al. 2009; Sucher et al. 2012). Recent studies reported chloroplast DNA barcoding markers in C. sativa that were informative for biogeographical origin and crop type prediction (Roman et al. 2019; Roman and Houston 2020). Since these regions represent the most highly polymorphic regions of the C. sativa chloroplast genome, they are referred to as “hotspots.” The polymorphisms were genotyped using CE-based methods, and Sanger sequencing revealed isoalleles at several loci with different repeat sequences or variations in the flanking regions that were not detected by CE (Roman et al. 2019; Roman and Houston 2020). This study seeks to expand upon previous studies by incorporating the “hotspot” barcoding regions into an MPS assay to provide more discriminatory results and a high throughput method for building a database of C. sativa chloroplast haplotypes.

Full chloroplast genome sequences have been reported for several marijuana and hemp cultivars (Matielo et al. 2020; Oh et al. 2016; Vergara et al. 2016). The full genome is 153,871 bp (Carmagnola and Dagestani cultivars) and contains 83 genes (Vergara et al. 2016). In comparison, the human mitochondrial genome is 16,569 bp (Anderson et al. 1981) (about a tenth of the size), and typically mitochondrial DNA analyses only involve sequencing a portion of the genome, usually the hypervariable regions (HV1 and HV2) (Ingman and Gyllensten 2006; Miller and Budowle 2001). Due to the large size of the C. sativa chloroplast genome, sequencing targeted regions (barcoding markers) gives better coverage and increased throughput capabilities compared to whole genome sequencing. The chloroplast genome is AT-rich (63%) and contains numerous homopolymeric stretches (Vergara et al. 2016). The MiSeq FGx® platform was chosen for sequencing in this study because has been shown to have a higher fidelity when sequencing homopolymeric stretches of DNA (Duke et al. 2015; Loman et al. 2012; McElhoe et al. 2014), and many of the polymorphisms identified in the previous studies (Roman et al. 2019; Roman and Houston 2020; Cheng and Houston, 2021) were homopolymeric STRs (hSTRs).

This study seeks to design an MPS panel for the MiSeq FGx® consisting of seven highly polymorphic “hotspot” regions in the C. sativa chloroplast (trnK-matK-trnK, rps16, trnS-trnG, ycf3, accD-psaI, clpP, and rpl32-trnL) to provide additional sequence data, discover isoalleles, and provide a high throughput method for creating a haplotype DNA database for hemp and marijuana samples. This assay could provide important investigative leads for law enforcement agencies investigating marijuana trafficking into and within the USA.