Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 May 25;18(2):e70031. doi: 10.1002/tpg2.70031

Discovery of major QTL and a massive haplotype associated with cannabinoid biosynthesis in drug‐type Cannabis

Maxime de Ronne 1,2,3, Davoud Torkamaneh 1,2,3,4,✉
PMCID: PMC12104491  PMID: 40415170

Abstract

Cannabis (Cannabis sativa L.), once sidelined by decades of prohibition, has now gained recognition as a multifaceted and promising plant in both medical research and commercial applications following its recent legalization. This study leverages a genome‐wide association study (GWAS) on 174 drug‐type Cannabis accessions from the legal Canadian market, focusing on identifying quantitative trait loci (QTL) and candidate genes associated with eleven cannabinoid traits using 282K common single‐nucleotide polymorphisms. This approach aims to transform our understanding of Cannabis genetics. We have pinpointed 33 significant markers that significantly influence cannabinoid production, promising to drive the development of Cannabis varieties with specific cannabinoid profiles. Among the notable findings is a massive haplotype of ∼60 Mb on chromosome 7 in Type I (i.e., tetrahydrocannabinol [THC]‐dominant) accessions, highlighting a major genetic influence on cannabinoid profiles. These insights offer valuable guidance for Cannabis breeding programs, enabling the use of precise genetic markers to select and refine promising Cannabis varieties. This approach promises to speed up the breeding process, reduce costs significantly compared to traditional methods, and ensure that the resulting Cannabis varieties are optimized for specific medical and recreational needs. This study marks a significant stride toward fully integrating Cannabis into modern agricultural practices and genetic research, paving the way for future innovations.

Core Ideas

  • Leveraging HD‐GBS and a high‐quality reference genome streamlines GWAS for uncovering complex traits in Cannabis breeding.

  • A total of 33 markers significantly associated with 11 cannabinoid traits were identified across the genome of drug‐type Cannabis.

  • A massive haplotype of ∼60 Mb and specific to Type I accessions was identified on chromosome 7.

Abbreviations

BIC

Bayesian information criterion

BLINK

Bayesian‐information and linkage‐disequilibrium iteratively nested keyway

CBC

cannabichromene

CBCA

cannabichromenic acid

CBCAS

cannabichromenic acid synthase

CBD

cannabidiol

CBDA

cannabidiolic acid

CBDAS

cannabidiolic acid synthase

CBDV

cannabidivarin

CBDVA

cannabidivarinic acid

CBG

cannabigerol

CBGA

cannabigerolic acid

CBGAS

cannabigerolic acid synthase

CBGVA

cannabigerovarinic acid

CBN

cannabinol

CBNA

cannabinolic acid

DAPC

discriminant analyses of principal components

ECS

endocannabinoid system

GO

gene ontology

GPP

geranyl diphosphate

GS

genomic selection

GTF

general feature format

GWAS

genome‐wide association study

HB

haplotype block

HD‐GBS

high‐density GBS

KASP

kompetitive allele specific PCR

LD

linkage disequilibrium

MAF

minor allele frequency

MAS

marker‐assisted selection

MEP

methylerythritol 4‐phosphate pathway

OA

olivetolic acid

PC

principal component

PT4

geranylpyrophosphate:olivetolate geranyltransferase 4

PVE

phenotypic variance explained

QQ

quantile–quantile

QTL

quantitative trait loci

RRS

reduced‐representation sequencing

SV

structural variations

THCA

Δ9‐tetrahydrocannabinolic acid

THCAS

tetrahydrocannabinolic acid synthase

THCV

tetrahydrocannabivarin

UGPMA

unweighted pair group method with arithmetic mean

WGS

whole‐genome sequencing

XP‐GBWAS

extreme‐phenotype GWAS

Δ8‐THC

Δ8‐tetrahydrocannabinol

Δ9‐THC

Δ9‐tetrahydrocannabinol

1. INTRODUCTION

Cannabis sativa L., commonly referred to as Cannabis, was domesticated around 8000 BCE, with early records suggesting Central Asia as its origin (Hurgobin et al., 2021). Throughout the 20th century, most countries imposed strict prohibitions on its cultivation and use (Bewley‐Taylor & Jelsma, 2011) before it was reintroduced in Canada as a significant source of seed oil and fiber in 1998, for medicinal products in 2001, and for recreational uses in 2018 (Cox, 2018; Welling et al., 2016). Initially cultivated as a multipurpose crop (Bai et al., 2021), selective breeding over centuries has led to the development of fiber‐ and drug‐producing cultivars, resulting in modern hemp and drug‐type (marijuana) populations (Clarke & Merlin, 2016; Kovalchuk et al., 2020). In just 4 years, drug‐type Cannabis, primarily cultivated for its medicinal and recreational phytocannabinoids, has emerged as one of Canada's most valuable agricultural sectors, contributing approximately $43.5 billion to the economy and creating over 150,000 jobs (Competition Bureau Canada, 2023).

Phytocannabinoids are meroterpenoids with a resorcinyl core with a para‐positioned isoprenyl, alkyl, or aralkyl side chain that were first isolated from Cannabaceae and have been recently found in Rhododendron species, some Fabaceae, Radula genus, and some fungi (Gülck & Møller, 2020). These compounds interact with the endocannabinoid system (ECS), a complex cell‐signaling system involved in regulating various physiological (e.g., satiety, pain, and energy metabolism) and psychiatric processes (e.g., schizophrenia) (Storr & Sharkey, 2007; Zou & Kumar, 2018). Phytocannabinoids can bind to or modulate the activity of ECS receptors (i.e., CB1 and CB2) and influence the synthesis and degradation of endogenous cannabinoids, thereby altering ECS activity (Mechoulam & Parker, 2013; Pertwee, 2008). Recent research highlights the therapeutic potential of phytocannabinoids for treating various conditions, including chronic pain, inflammation, depression, and epilepsy (Lu & MacKie, 2016; Russo, 2016; Zou & Kumar, 2018), making the Cannabis plant particularly attractive for the pharmaceutical industry.

Despite Cannabis producing over 545 potentially bioactive secondary metabolites (Martínez et al., 2020) and over 120 cannabinoids (Hanuš et al., 2016; Rock & Parker, 2021), only a few, such as Δ9‐tetrahydrocannabinolic acid (THCA), cannabidiolic acid (CBDA), and their precursor cannabigerolic acid (CBGA), are found in significant amounts. This has led to the classification of Cannabis cultivars into distinct chemotypes based on cannabinoid profiles (Hanuš et al., 2016): Type I cultivars predominantly produce THCA, type II maintain a balance between THCA and CBDA, type III are rich in CBDA, type IV are characterized by higher levels of CBGA, and type V cultivars are essentially cannabinoid‐free.

The biosynthesis of cannabinoids (Figure 1), primarily occurring within the trichomes of female flowers, results in the production of mainly acidic cannabinoids (Gülck & Møller, 2020; Luo et al., 2019; Sirikantaramas et al., 2005). The synthesis begins in the cytosol, where enzymes catalyze the formation of olivetolic acid (OA) and divarinic acid (DA) from fatty acids and isoprenoid precursors via the polyketide pathway (Gagne et al., 2012). In parallel, in the plastidial organelles, typically chloroplasts, the geranyl diphosphate (GPP) is synthesized through the methylerythritol 4‐phosphate (MEP) pathway. Geranylpyrophosphate:olivetolate geranyltransferase 4 (PT4), also known as cannabigerolic acid synthase (CBGAS), catalyzes the prenylation of OA and DA with GPP, forming the CBGA precursors and cannabigerovarinic acid (CBGVA), respectively (Fellermeier et al., 2001; Luo et al., 2019). Following synthesis, the cannabinoids undergo oxidative cyclization within the resin cavities of the trichomes, a process facilitated by several oxidocyclases (Geissler et al., 2018; Rodziewicz et al., 2019). These enzymes include cannabichromenic acid synthase (CBCAS), cannabidiolic acid synthase (CBDAS), and tetrahydrocannabinolic acid synthase (THCAS), which convert CBGA to cannabichromenic acid (CBCA), CBDA, cannabidivarinic acid (CBDVA), THCA, and Δ9‐tetrahydrocannabivarinic acid (THCVA), respectively (Kim et al., 2022; Morimoto et al., 1997; Sirikantaramas et al., 2004; Taura et al., 2007). Cannabinoids are predominantly stored in their acidic forms within the plant but can convert to their neutral forms, such as cannabigerol (CBG), cannabichromene (CBC), cannabidiol (CBD), Δ9‐tetrahydrocannabinol (Δ9‐THC), cannabinol (CBN), cannabidivarin (CBDV), and tetrahydrocannabivarin (THCV), through nonenzymatic thermal decarboxylation when exposed to light, heat, or combustion (Tan et al., 2018). Additionally, cannabinoids can undergo spontaneous chemical changes, including oxidation to cannabinolic acid (CBNA) and isomerization to Δ8‐tetrahydrocannabinol (Δ8‐THC).

FIGURE 1.

FIGURE 1

Open in a new tab

Biosynthesis pathway of cannabinoids in Cannabis. Enzymes are in italics: geranylpyrophosphate:olivetolate geranyltransferase 4 (PT4); cannabigerolic acid synthase (CBGAS); cannabichromenic acid synthase (CBCAS); cannabidiolic acid synthase (CBDAS); and tetrahydrocannabinolic acid synthase (THCAS). Cannabinoid precursor: methylerythritol 4‐phosphate (MEP); geranyl diphosphate (GPP); cannabigerolic acid (CBGA); and cannabigerovarinic acid (CBGVA). Acidic cannabinoids: cannabichromenic acid (CBCA); cannabidiolic acid (CBDA); Δ9‐tetrahydrocannabinolic acid (THCA); cannabinolic acid (CBNA); cannabidivarinic acid (CBDVA); and Δ9‐tetrahydrocannabivarinic acid (THCVA). Neutral cannabinoids: cannabigerol (CBG); cannabichromene (CBC); cannabidiol (CBD); Δ9‐tetrahydrocannabinol (Δ9‐THC); Δ8‐tetrahydrocannabinol (Δ8‐THC); cannabinol (CBN); cannabidivarin (CBDV); and tetrahydrocannabivarin (THCV). The cannabinoids measured for this study are highlighted in purple.

During the twentieth century, prohibition led to clandestine Cannabis breeding that relied on undocumented methods and lacked access to modern technologies (Welling et al., 2016). The primary focus of these efforts was to enhance sexual characteristics such as the production of female flowers and increasing cannabinoids, particularly THC. This selective breeding favored a broad diversity of traits associated with drug‐type Cannabis over those related to hemp (Clarke & Merlin, 2016; Welling et al., 2016). While these practices expanded the phenotypic and chemotypic variation, they also contributed to the loss of many varieties due to the war on drugs (Torkamaneh & Jones, 2021). Like other high‐value crops, Cannabis stands to benefit significantly from modern breeding technologies such as marker‐assisted selection (MAS; Collard & Mackill, 2008; Francia et al., 2005) and genomic selection (GS) (de Koning, 2016; Jannink et al., 2010). These technologies, which have been successful in other crops, offer the potential to enhance desired traits ranging from improved fiber and seed production for industrial uses to specific chemotype profiles for medicinal and recreational purposes (Hesami et al., 2020; Morrell et al., 2012). Therefore, a deep understanding of the genetic basis of cannabinoid traits in drug‐type Cannabis is a prerequisite for developing improved Cannabis varieties with specific cannabinoid profiles and concentrations.

Recent legislative changes have significantly improved access to comprehensive genomic and transcriptomic datasets, simultaneously simplifying the regulatory framework for both agronomic and medical Cannabis research (Grassa et al., 2021; Hesami et al., 2020; Monthony et al., 2024). These changes have facilitated various genetic studies, such as quantitative trait loci (QTL) mapping on biparental populations, which have identified maturity‐related QTL in hemp (Bakker et al., 2021). In parallel, genome‐wide association studies (GWAS) on large admixed populations have uncovered promising QTL associated with fiber quality and flowering (Petit, Salentijn, Paulo, Denneboom, & Trindade, 2020; Petit, Salentijn, Paulo, Denneboom, van Loo, et al., 2020), as well as salt tolerance (Sun et al., 2023) in hemp and terpenes in drug‐type Cannabis (Watts et al., 2021). Additionally, an extreme‐phenotype GWAS (XP‐GWAS) approach using whole‐genome sequencing (WGS) data from a bulked sample analysis (BSA) panel of 711 individuals from 72 accessions, diverging for CBDA and THCVA concentrations, has identified key molecular markers and candidate genes (Welling et al., 2020). Finally, the use of a chromosome‐scale reference genome (i.e., Cs10) (Grassa et al., 2021) combined with several hundred thousand single‐nucleotide polymorphisms (SNPs) from high‐density genotyping‐by‐sequencing (HD‐GBS) data (Torkamaneh et al., 2021) has enabled the identification of multiple QTL associated with nine agronomic and morphological traits in drug‐type Cannabis (de Ronne et al., 2024).

Core Ideas

  • Leveraging HD‐GBS and a high‐quality reference genome streamlines GWAS for uncovering complex traits in Cannabis breeding.

  • A total of 33 markers significantly associated with 11 cannabinoid traits were identified across the genome of drug‐type Cannabis.

  • A massive haplotype of ∼60 Mb and specific to Type I accessions was identified on chromosome 7.

In this study, we performed GWAS focusing on eleven cannabinoid traits. We successfully identified major QTL, putative candidate genes, and putative causal SNPs associated with high variation in cannabinoid content. One of the notable findings was a massive haplotype of ∼60 Mb on chromosome 7, primarily found in type I accessions. This massive haplotype overlaps with the canonical THCAS and CBDAS on chromosome 7 and could lead to chemotypic differentiation in the GWAS panel. Further validation is required to confirm the extent, stability, and functional impact of this haplotype, particularly across larger and more chemically diverse Cannabis populations. Identifying the genetic basis of desired traits is essential for modern breeding programs that utilize MAS and GS. This knowledge significantly contributes to the development of improved and specifically tailored Cannabis cultivars, thereby enhancing the efficiency and effectiveness of breeding efforts.

2. MATERIALS AND METHODS

2.1. Plant material

All research activities, including the acquisition and cultivation of cannabis plants, were carried out under our cannabis research license (LIC‐QX0ZJC7SIP‐2021) and in full adherence to Health Canada's regulations. A panel of 174 drug‐type accessions, previously phenotyped by Lapierre et al. (2023), was used in this study (Table S1). These accessions were carefully selected from various sources to provide a comprehensive representation of the drug‐type cannabis varieties available in the legal Canadian market.

2.2. Cannabinoid profiling

Biochemical analysis of trimmed and dried flowers was conducted at the Metabolomics Platform in the Institute of Nutrition and Functional Foods, Université Laval, Québec, QC, Canada, as outlined by Lapierre et al. (2023). In this study, we analyzed the concentration (% w/w) of 10 cannabinoids: CBGA, THCA, CBDA, Δ9‐THC, CBD, CBG, CBC, CBN, CBDV, and THCV. The accessions were classified according to the proportion of the two main cannabinoids, that is, the estimated THC and CBD potential compared through THC/CBDratio:THC potential/(THC potential + CBD potential). The estimated THC potential is calculated based on the concentration of THCA, which could be converted to Δ9‐THC, where M is the molar mass of Δ9‐THC (314.5 g/mol) and THCA (358.5 g/mol), and the factor 0.877 is derived from the molar mass of Δ9‐THC divided by the molar mass of THCA: THC potential = M Δ9‐THC + 0.877 × M THCA. Similarly, the CBD potential is calculated as follows: CBD potential = M CBD + 0.877 × M CBDA. Histograms representing the distribution of each trait for the 174 accessions were generated using R v4.2.1 (R Core Team, 2021) with the “hist” function.

2.3. High‐density genotyping

The sequencing and genotyping of the GWAS panel was performed using HD‐GBS as described in de Ronne et al. (2024). Briefly, DNA was extracted from approximately 50 mg of young leaf tissue from each accession using the CTAB‐chloroform protocol (Aboul‐Maaty & Oraby, 2019). DNA quantification was carried out using a Qubit fluorometer with the dsDNA HS assay kit (Thermo Fisher Scientific), and concentrations were adjusted to 10 ng/µL for all samples. Final DNA samples were used to prepare HD‐GBS libraries with the BfaI enzyme as described in Torkamaneh et al. (2021) at the Institut de biologie intégrative et des systèmes, Université Laval, QC, Canada. Sequencing was conducted on an Illumina NovaSeq 6000 (Illumina) with 150 paired‐end reads at the Genome Quebec Service and Expertise Center, Montreal, QC, Canada.

Sequencing data were processed with the Fast‐GBS v2.0 (Torkamaneh et al., 2020) using the C. sativa cs10 v2 reference genome (GenBank acc. no. GCA_900626175.2; Grassa et al., 2018). Raw SNP data were filtered with VCFtools (Danecek et al., 2011) to remove low‐quality SNPs (QUAL < 10 and MQ < 30) and variants with a proportion of missing data exceeding 80%. Missing data imputation was performed with BEAGLE 4.1 (Browning & Browning, 2016), followed by a second round of filtration, retaining only biallelic variants with heterozygosity less than 50% and a minor allele frequency (MAF) more than 6%. Additionally, variants residing on unassembled scaffolds were removed. The resulting catalog of 281,795 SNPs was used to conduct population structure analysis and GWAS. The proportion of heterozygous genotypes between samples was estimated using TASSEL5 (Bradbury et al., 2007). The nucleotide diversity (π; Nei & Li, 1979) was measured in a sliding window of 1000 bp every 500 bp across the genome using –window‐pi and –window‐pi‐step options of VCFtools (Danecek et al., 2011).

2.4. Population structure analysis

Population structure and accession admixture were determined using a variational Bayesian inference algorithm implemented in fastStructure (Raj et al., 2014) for a number of subpopulations (K) set from 1 to 10. The optimal number of K (i.e., K = 3) explaining the population complexity was estimated using the ChooseK tool from fastStructure, and admixture proportions were visualized using Distruct v2.3. The kinship matrix (K*; Figure S1) was generated using TASSEL5 (Bradbury et al., 2007) with the Centered_IBS method and plotted with GAPIT v3 (J. Wang & Zhang, 2021).

Population structure was further investigated using DAPC (Jombart et al., 2010) using the R package “adegenet.” The number of clusters was estimated using the “find.cluster” function with a maximum limit set to 40 clusters and 200 PCs. The optimal number of clusters (i.e., K = 3) was determined by the Bayesian information criterion (BIC) value for different numbers of K (Figure S2a,b). To visualize the DAPC using the “scatter” function, the optimal number of PCs was estimated with two cross‐validation procedures using “optim.a.score” (i.e., PCs = 21, Figure S2c) and “xvalDapc” (i.e., PCs ≤ 20, Figure S2d). Cluster assignments from both fastStructure and DAPC were compared using the “table” function for a K value of 3. The phylogenetic tree was estimated with the unweighted pair group method with arithmetic mean (UPGMA) and constructed using the R package “phytools,” with each branch colored according to the clustering results from fastStructure.

2.5. Genome‐wide association analysis

Marker‐trait association analysis was performed with the Bayesian‐information and linkage‐disequilibrium iteratively nested keyway (BLINK) using GAPIT v3 (J. Wang & Zhang, 2021), using the ∼282K high‐quality SNPs and the phenotyping data for the 10 cannabinoids and THC/CBDratio (eleven traits in total). The identification of false positive was minimized by incorporating population structure (i.e., P matrix generated with fastStructure for K = 3) and kinship (i.e., K* matrix generated with TASSEL5). The threshold of significance for marker‐trait associations (i.e., 1.70 × 10−7) was set to ensure a false discovery rate < 0.05, adjusted with a Benjamini‐Hochberg correction (Benjamini & Hochberg, 1995). Markers with a phenotypic variance explained (PVE) less than 3% were excluded from the analysis as they were considered uninformative and of limited interest to breeders. Manhattan plots showing −log10(p) distribution of markers by chromosome were generated with rMVP (Yin et al., 2021) using plot.type = “m”, and QQ plots were created with GAPIT v3 (J. Wang & Zhang, 2021). Boxplots of the allelic classes of significant markers were generated with “ggplot2” in R.

2.6. Candidate gene identification

Pairwise‐linkage disequilibrium (LD) was calculated with PLINK v1.9 (Purcell et al., 2007) using –r2 –ld‐window‐r2 0 parameters. Markers in high LD (r 2 ≥ 0.75) with significant markers were retained to define haplotype blocks (HBs). Markers failing to form HB and residing outside of genetic regions were removed from the candidate gene investigation. Genes with significant SNPs in their sequence or located in HBs (defined by the 5′‐most and 3′‐most marker of the HB) were considered as candidate genes. The gene ontology (GO) annotations of these candidate genes were examined based on the description provided by the NCBI Cannabis sativa Annotation Release 100 (cs10; NCBI, 2019). To further confirm and provide a more detailed functional annotation of candidate genes, phylogenetic ortholog inferences were performed using OrthoFinder (Emms & Kelly, 2019) with the Arabidopsis thaliana transcriptome (TAIR 11; Cheng et al., 2017) as described in Carey et al. (2021). The functional impact of significant SNPs or SNPs residing in HBs was investigated using the Ensembl Variant Effect Predictor (McLaren et al., 2016) with the general feature format (GTF) of cs10 reference genome. Among these SNPs, only those located in candidate gene sequences were retained. Consequence terms (e.g., in‐frame insertion, missense variant, and synonymous variant) and the impact rating are based on Sequence Ontology (Eilbeck et al., 2005).

3. RESULTS

3.1. Cannabinoid profiling and chemotype in the GWAS panel

The GWAS panel exhibited a wide range of cannabinoid profiles and concentrations (Figure 2). The THCA and CBDA were the primary cannabinoids measured in terms of concentration, with CBGA, their precursor, being the third major compound (Figure 2a). All other compounds derived from CBGA, THCA, and CBDA were present in very low concentrations (<0.8% w/w). No Δ8‐THC was detected in this population. The concentrations of CBN, THCV, and CBDV were not quantifiable and were therefore treated qualitatively (i.e., present/absent). Considering the THC and CBD potential via the THC/CBDratio, the GWAS panel clearly delineated three categories from 4.4% to 5.9%, 23.2% to 58.0%, and 96.7% to 99.8% of THC/CBDratio (Figure 2b), respectively categorized in Type III (n = 5 CBD‐dominant), Type II (n = 17 THC/CBD intermediate), and Type I (n = 152 THC‐dominant) chemotypes. Type I accessions mainly produced THCA (μ = 22.3% w/w) compared to Type II accessions (μ = 8.0% w/w) and Type III accessions (μ = 0.7% w/w). Conversely, CBDA was almost absent in Type I accession (μ = 0.05% w/w) compared to Type II (μ = 9.1% w/w) and Type III accessions (μ = 14.5% w/w). A similar pattern was observed with CBC, which was very low in Type I accessions (μ = 0.03% w/w) compared to Type II accessions (μ = 0.14% w/w) and Type III accessions (μ = 0.24% w/w). The Δ9‐THC was specific and produced equally by Type I and Type II (THC‐producing) accessions. While the THCV was identified in the three chemotypes, CBDV and CBD were specific to the Type II and Type III (CBD‐producing) accessions. The CBN was specific to the Type I accessions. Although Type I accessions produced more THCA than Type II accessions, both types exhibited a similar residual amount of Δ9‐THC. This observation is likely due to the passive conversion of THCA into Δ9‐THC via nonenzymatic thermal decarboxylation (Tan et al., 2018), suggesting that both groups undergo comparable levels of THCA decarboxylation.

FIGURE 2.

FIGURE 2

Open in a new tab

Cannabinoid profiling of 174 drug‐type Cannabis accessions. (a) Frequency distribution for the eleven cannabinoid traits. The weight/weight percentage (% w/w) corresponds to the weight proportion of cannabinoids relative to the total weight of the sample analyzed. Cannabinoid data are detailed in Table S1. (b) Heatmap of the cannabinoid concentration and presence in the genome‐wide association study (GWAS) panel. Each vertical line represents an accession. The color graduation represents the normalized concentration (min–max scaling) for each trait independently. The accessions are sorted from left to right, from the highest to the lowest tetrahydrocannabinol (THC)/cannabidiol (CBD)ratio. The GWAS panel comprised 152 Type I, 17 Type II, and five Type III accessions. CBC, cannabichromene; CBD, cannabidiol; CBDA, cannabidiolic acid; CBDV, cannabidivarin; CBG, cannabigerol; CBGA, cannabigerolic acid; CBN, cannabinol; THC, tetrahydrocannabinol; THCA, tetrahydrocannabinolic acid; THCV, tetrahydrocannabivarin.

3.2. Genotyping and population structure analysis

The HD‐GBS provided a dense catalog of ∼282K high‐quality SNPs well distributed across the genome. On average, there was one marker per ∼3 kb, with only 12 gaps exceeding 1 Mb, the largest being 1.2 Mb (Figure 3). Within this catalog, 25.5% of the genotypes were heterozygous, with similar proportions observed between chromosomes and Cannabis Types. In contrast, the proportion of heterozygous genotypes ranged from 13.8% to 45.9% between accessions. Genome‐wide nucleotide diversity (θπ) in the GWAS panel was 1.05 × 10−3, with relatively homogeneous values across autosomes (9.75 × 10−4 to 1.15 × 10−3), while chromosome X exhibited the lowest diversity (8.66 × 10−4; Table S2). Nucleotide diversity among Cannabis Types was relatively similar, ranging from 1.05 × 10−3 to 9.94 × 10−4. Chromosome‐specific trends mirrored these patterns, except for chromosome 3 of Type II accessions exhibiting reduced diversity across all chromosomes. There was no correlation observed between chemotypes and proportion of heterozygous genotypes or nucleotide diversity.

FIGURE 3.

FIGURE 3

Open in a new tab

Genome‐wide association studies of eleven cannabinoid traits in drug‐type Cannabis and candidate gene identification. (a) Manhattan plot showing significant associated single‐nucleotide polymorphisms (SNPs). The colors of the heatmap correspond to the number of SNPs within 1 Mb windows. (b) Visual representation of the SNP_23 haplotype block (HB) and a putative causal SNP residing in LOC115696275 gene. The SNP_23 is highlighted in red. The SNP_23 formed a strong HB with eight SNPs. The value represents pairwise linkage disequilibrium (LD) (r 2 × 100), and the empty square indicates complete LD (r 2 = 1). (c) Distribution of SNPs residing in SNP_23 HB on chromosome 7 of cs10. The putative causal SNP at Chr07:6,403,827 was indicated by a dashed vertical line. The positions of SNPs 1 through 9 were located on chromosome 7 at the following base pair coordinates: 6,403,713; 6,403,717; 6,403,726; 6,403,732; 6,403,736; 6,403,746; 6,403,813; 6,403,827; and 6,403,915 bp. (d) Average tetrahydrocannabinol (THC) and cannabidiol (CBD) potential compared to allelic classes of the putative causal SNP at Chr07:6,403,827. CBC, cannabichromene; CBDA, cannabidiolic acid; CBDV, cannabidivarin; CBG, cannabigerol; CBGA, cannabigerolic acid; CBN, cannabinol; THCA, tetrahydrocannabinolic acid; THCV, tetrahydrocannabivarin.

To minimize potential confounding effects in the GWAS analysis, we assessed both population structure (i.e., P matrix) and kinship (i.e., K* matrix), as these factors can lead to biases in allele frequency distributions and spurious associations. The estimation of admixture among accessions indicated that three principal components (PCs) were optimal for explaining the structure of the population (Figure S3a). This finding was supported by discriminant analyses of principal components (DAPC), which showed the minimal BIC value at K = 3 (Figures S2b and S3b). Both fastStructure and DAPC consistently identified K = 3 as the optimal value, with 93.6% concordant assignment (Figure S3d). Phylogenetic trees and relatedness analysis revealed low intra‐ and intercluster genetic diversity, with accessions appearing neither significantly similar nor significantly distant (Figures S1 and S3c). Among population structure clusters, K1 and K2 showed similar genome‐wide diversity (1.12 × 10−3 and 1.20 × 10−3), while K3 had a slightly lower value (8.44 × 10−4; Table S2). Moreover, there was no correlation observed between clustering assignments and heterozygous genotypes, nucleotide diversity, cannabinoid traits, chemotypes, or germplasm origins.

3.3. Genome‐wide association analysis of cannabinoids

To improve the detection of true associations in a dense SNP catalog and for complex traits like cannabinoid content, a multi‐locus model, BLINK, was used to perform marker‐trait association analysis (Figure 3). The quantile–quantile (QQ) plots show that the majority of observed p‐values align closely with the expected distribution under the null hypothesis, indicating a well‐fitted model (Figure S4). This suggests that the P and K* matrices effectively controlled for population structure. A total of 33 markers were identified as significantly associated with the 11 cannabinoid traits (FDR p‐value < 1.70 × 10−7; Table 1; Figure 3a). Among them, 19 were associated with major QTL, that is, exceeding 10% of PVE in the GWAS panel, with at least one for each trait. Notably, SNP_1 and SNP_12 demonstrated exceptionally high PVE, with values of 96.0% and 89.1%, respectively. Chromosome 7 carried the markers with the highest association score (−log10(p) > 25). SNP_12 and SNP_21 were associated with several correlated traits, including THC/CBDratio and CBD, or CBD and CBDA. For instance, the “T” allele of SNP_12 (Chr07:18,615,635) was linked to a higher THC/CBDratio, indicating greater THC potential, while the “G” allele was associated with higher CBD and CBC concentration (Figure 3a; Figure S5). The SNP_21 (Chr08:47,102,284) was associated with a minor QTL for CBD and a major QTL for CBDA. As shown in Figure S5, for all markers studied, no evidence of heterosis was observed for these traits.

TABLE 1.

List of markers associated with cannabinoid traits in drug‐type Cannabis by genome‐wide association studies (GWASs).

Traits Marker ID Chr. MSS position Major/minor allele MAF (%) p‐value PVE (%) Effect a
CBGA SNP_1 9 59648694 C/T 8.9 5.66E‐08 96.0 −0.41
THCA SNP_2 1 36547962 T/A 17.2 1.82E‐09 8.5 3.63
SNP_3 5 17239444 A/T 10.9 1.30E‐08 16.8 3.98
SNP_4 6 30634679 T/C 18.4 1.47E‐09 16.7 3.43
SNP_5 7 18626700 C/T 6.9 5.34E‐08 32.4 5.31
SNP_6 X 16559033 A/G 33.0 1.37E‐10 8.7 3.28
Δ9‐THC SNP_7 2 3864032 G/A 6.6 5.99E‐15 76.7 −0.10
SNP_8 6 31860692 C/T 18.4 2.26E‐18 8.8 0.07
THC/CBDratio SNP_9 4 55987941 T/C 8.0 5.03E‐10 19.7 −0.06
SNP_10 4 83600878 C/T 8.9 3.30E‐09 5.4 0.05
SNP_11 6 30464747 C/G 10.6 2.22E‐14 9.1 0.07
SNP_12 7 18615635 T/G 6.9 2.91E‐59 51.6 0.36
SNP_13 7 59374763 C/T 6.9 2.45E‐16 12.2 0.12
THCV SNP_14 2 25985421 A/G 10.3 5.00E‐09 9.2 0.23
SNP_15 3 16999237 A/G 14.1 1.13E‐08 28.6 0.19
SNP_16 6 6672721 C/A 31.0 3.62E‐09 7.8 −0.14
SNP_17 6 72169351 C/T 13.8 6.28E‐09 30.8 0.20
CBDA SNP_18 7 44687324 T/C 8.9 1.27E‐43 18.6 −4.49
SNP_19 7 66722326 G/A 6.0 1.16E‐27 14.7 −3.13
SNP_20 8 47100871 G/C 6.6 3.55E‐09 15.3 −4.56
SNP_21 8 47102284 C/T 7.2 5.63E‐12 37.9 −1.04
CBD SNP_22 6 76796257 T/A 7.8 3.16E‐10 13.0 −0.01
SNP_23 7 6403732 C/‐ 10.1 6.65E‐09 6.2 −0.01
SNP_12 7 18615635 T/G 6.9 2.27E‐40 66.0 −0.06
SNP_24 8 41807341 A/G 7.8 5.87E‐11 3.7 0.02
SNP_21 8 47102284 C/T 7.2 1.42E‐19 5.5 −0.03
CBG SNP_25 1 10081692 T/‐ 34.8 1.02E‐07 6.8 0.01
SNP_26 3 66045720 T/G 6.3 1.48E‐07 56.7 −0.03
SNP_27 6 49801371 A/‐ 42.5 1.04E‐13 7.4 −0.02
CBC SNP_12 7 18615635 T/G 6.9 6.16E‐31 89.1 −0.07
CBN SNP_28 5 4807613 A/G 6.3 2.75E‐12 53.4 0.10
SNP_29 X 66914884 A/G 8.3 8.11E‐20 35.6 0.14
SNP_30 X 76104119 C/A 8.3 1.18E‐07 5.1 0.06
CBDV SNP_31 4 54151119 C/T 12.1 1.10E‐07 32.4 0.52
SNP_32 7 34814767 C/A 7.5 5.30E‐08 6.8 0.09
SNP_33 7 55724129 T/A 7.5 5.18E‐09 5.1 0.14
Open in a new tab

Note: Marker‐trait associations were performed with the Bayesian‐information and linkage‐disequilibrium iteratively nested keyway (BLINK) method.

Abbreviations: CBC, cannabichromene; CBD, cannabidiol; CBDA, cannabidiolic acid; CBDV, cannabidivarin; CBG, cannabigerol; CBGA, cannabigerolic acid; CBN, cannabinol; MAF: minor allele frequency; MSS, most significant SNPs; PVE, phenotypic variance explained; SNP, single‐nucleotide polymorphism; THC, tetrahydrocannabinol; THCA, tetrahydrocannabinolic acid; THCV, tetrahydrocannabivarin.

a

Effect referred to the major allele.

3.4. Identification of putative candidate genes

Due to the complex nature of cannabinoid modulation and the inherent limitation of GWAS resolution, determining whether an SNP or gene is truly causal is not possible without further functional characterization, regardless of the presence or absence of an obvious functional annotation. Nevertheless, significant SNPs are associated with phenotypic variation, suggesting an interaction with causal variants that may also have transcriptional, posttranscriptional, or functional effects. Therefore, genomic regions of interest for candidate gene discovery were defined by HBs composed of significant SNPs. While this approach may not definitively pinpoint causal variants, it narrows down genomic regions associated with significant SNPs, providing the potential to identify variants and genes that may co‐segregate with these markers and contribute to the traits of interest. Only a limited number of SNPs exhibited strong LD (r 2 ≥ 0.95). Consequently, genetic regions of interest were defined by HBs composed of SNPs in high LD (r 2 ≥ 0.75) with a significant SNP. Among the 34 significant SNPs, 23 formed 21 HBs ranging from 2 to 20 markers and spanning 70 bp to 121.2 kb (Table S3). Both SNP_5 and SNP_12 were present within the same HB, as well as SNP_20 and SNP_21. Eleven significant SNPs were not in high LD with others and did not form haplotype blocks.

Thirteen candidate genes were identified under eight HB regions of interest, with six from the SNP_10′s HB alone (Table S3). Additionally, five significant SNPs (SNP_6, SNP_7, SNP_15, SNP_25, and SNP_26) that did not form HBs were located within different gene sequences, resulting in a total of 18 candidate genes. Among them, seven were of unknown function in the cs10 genome annotation. Orthological analysis by aligning the protein sequences of candidate genes with those of the Arabidopsis proteome revealed that for most candidate genes and their corresponding orthologs, functional annotations were found to be consistent. In cases where the function of a gene was unknown, Arabidopsis orthologs with known functions were considered (Table S3). The three most represented functional annotations among the candidate genes were related to enzymes (LOC115706556, LOC115718629, LOC115714462, LOC115714401, LOC115725486, and LOC115725486), transcription factors (LOC115711687, LOC115714441, LOC115725341, and LOC115724958), and non‐LTR retroelement reverse transcriptases (LOC115713787 and LOC115695103).

A total of 113 SNPs (i.e., 33 significant SNPs and 80 residing in HBs of interest) were analyzed for their putative impact on gene function (Tables S3 and S4; Figure S6). While variants outside of candidate gene sequences can impact gene regulation by affecting cis‐ or distal‐regulatory elements, examining their potential functional effects through bioinformatics analysis alone is particularly challenging. In contrast, only SNPs located within candidate gene sequences were retained (Table S4), as they are more likely to directly impact gene function. No SNPs with a predicted high impact, such as frameshift variants or start/stop codon gains or losses, were identified. Four missense variants, each resulting in an amino acid substitution with distinct biochemical properties that potentially alter protein structure, were identified (Table S4). Additionally, one SNP located in the three prime untranslated region (UTR) was found, both of which were considered as identified as potentially altering gene functions (Table S4) and were further investigated for their impact on THC and CBD potential (Figure S6). The alternative allele (relative to the CBD‐dominant cs10 reference genome) “C” at Chr01:36,470,068 was in an HB with SNP_2 and caused a missense substitution in LOC115706595 (ribosome biogenesis protein NOP53), resulting in a recessive and beneficial impact on THC potential levels. The SNP_7 was in the 3′‐UTR of LOC115718629 (nicotinamidase 1) and at ∼300 bp downstream of LOC115718628 (chloroplast protein EXECUTER 1). The “A” allele was associated with a high and stable level of THC potential and absence of CBD, while the “G” allele was linked to a wide range of THC and CBD potentials. The alternative allele “C” at Chr06:31,864,145, linked with SNP_8, induced a missense substitution in LOC115695103 (non‐LTR retroelement reverse transcriptase), resulting in a recessive and reductive impact on THC potential. The alternative “T” allele of SNP_22 induced a missense substitution in LOC115695528 (receptor‐like protein EIX2), with accessions carrying this allele exhibiting, on average, twice the THC potential compared to those carrying the “A” allele. Finally, the SNP at Chr07:6,403,827 formed an HB with SNP_23 spanning ∼300 bp and present in two forms (Figure 3b). This SNP induced a missense substitution in LOC115696275 (Figure 3c; Table S4). The reference allele “C” and alternative allele “T” were specific to high CBD and THC potential, respectively, with a codominant effect (Figure S7). Accessions carrying the reference allele “T” (or Haplotype I) exhibited high THC potential and no CBD, while those carrying the alternative allele “C” (or Haplotype II) showed the opposite pattern (Figure 3d; Figure S7). Accessions carrying both alleles exhibited balanced THC and CBD concentrations.

3.5. Type I accessions carry a massive haplotype on chromosome 7

The distribution of significant SNPs across the three chemotype categories was investigated (Figure S7). The eight significant SNPs identified on chromosome 7 formed a haplotype composed of the major alleles present in over 87% of Type I accessions (Figure 4a,b). When considering only the SNP_5, SNP_12, SNP_13, SNP_18, SNP_19, and SNP_23, this haplotype was carried by approximately 95% of Type I accessions. The Type III accessions predominantly carried the minor alleles at these SNPs, without forming a complete haplotype, whereas the Type II accessions mainly exhibited heterozygosity. Notably, the major alleles of SNP_12 (Figure 4c) and SNP_19 (Figure 4d) were present in 100% of Type I accessions and were associated with high phenotypic variations.

FIGURE 4.

FIGURE 4

Open in a new tab

Six significant single‐nucleotide polymorphisms (SNPs) formed a massive haplotype related to Type I accessions on Chromosome 7. (a) Eight significant SNPs were identified on chromosome 7. (b) Allelic classes of significant SNPs compared to chemotypes. The SNPs were organized based on their positions along chromosome 7, from the beginning to the end of the chromosome. Each vertical line represents an accession. Blue, gray, and orange represent tetrahydrocannabinol (THC)‐related, heterozygous, and cannabidiol (CBD)‐related alleles, respectively. The genome‐wide association study (GWAS) panel comprised 152 Type I, 17 Type II, and five Type III accessions. (c) Distribution of allelic classes for significant association with cannabinoid traits of SNP_12 and SNP_19. CBC, cannabichromene; CBDA, cannabidiolic acid; CBDV, cannabidivarin; CBG, cannabigerol; CBGA, cannabigerolic acid; CBN, cannabinol; THCA, tetrahydrocannabinolic acid; THCV, tetrahydrocannabivarin.

4. DISCUSSION

Cannabis holds significant attractiveness for medical research, as well as pharmaceutical development and the recreational industry, due to its prolific production of cannabinoids (Baratta et al., 2022; Coelho et al., 2023; Gülck & Møller, 2020), which interact with the ECS, influencing a wide range of physiological and psychological processes (Russo, 2016; Storr & Sharkey, 2007; Zou & Kumar, 2018). Historically, years of prohibition have impeded the establishment of genetic resource collections and the development of advanced breeding practices, thus limiting both the genetic improvement and the understanding of Cannabis traits (Welling et al., 2016). In the context of expanding and diversifying uses of Cannabis, the development of tailored Cannabis varieties becomes crucial to meet the various needs and challenges encountered by research and industry (Torkamaneh & Jones, 2021). This necessitates genetic screening of promising germplasm and the implementation of modern breeding tools, such as MAS or GS, which rely on the identification of molecular markers associated with high‐value traits, including yield, resistance to biotic stresses, and cannabinoid production.

The present study builds on global efforts to explore genetic markers of interest in hemp and drug‐type Cannabis using marker‐trait association studies (Bakker et al., 2021; Petit, Salentijn, Paulo, Denneboom, and Trindade, 2020; Petit, Salentijn, Paulo, Denneboom, van Loo, et al., 2020; Sun et al., 2023; Welling et al., 2020; Watts et al., 2021). Utilizing the same genotyping dataset and GWAS procedure we previously employed to identify major QTL associated with agronomic and morphological traits in drug‐type Cannabis (de Ronne et al., 2024), we discovered major QTL associated with cannabinoid traits, including CBGA, CBDA, THCA, CBG, CBC, CBD, Δ9‐THC, CBN, CBDV, THCV, and THC/CBDratio. The distribution of the GWAS panel for the THC/CBDratio suggested a Mendelian inheritance model of chemotype that involved a single locus with codominant alleles, as proposed by the Bt :Bt (Type I, THCA/THCA), Bt :Bd (Type II, THCA/CBDA), and Bd :Bd (Type III, CBDA/CBDA) models (De Meijer et al., 2003). Two significant SNPs on chromosome 7, SNP_12 and SNP_13, supported this observation, with accessions categorized as Type I or Type III depending on their alleles and as Type II when heterozygous. Additionally, a third non‐synonymous SNP at Chr07:6,403,827, in very strong LD with SNP_23, caused a Glu to Lys substitution in an exon of a 2‐oxoglutarate (2OG) and Fe(II)‐dependent oxygenase superfamily protein, demonstrating a similar allele‐dependent chemotype pattern. This non‐synonymous SNP provides evidence of being a putative causal variant, modulating the qualitative production of THCA and CBDA through sequence variation in LOC115696275. Further functional characterization of LOC115696275 could confirm its implication in the biosynthetic pathway of THCA and CBDA with allele‐dependent chemotype effect. These three SNPs (SNP_12, SNP_13, and at Chr07:6,403,827) could be associated with qualitative production of THCA or CBDA and therefore could be used to develop chemotype detection assays such as kompetitive allele specific PCR (KASP; Semagn et al., 2014), TaqMan assays (De La Vega et al., 2005), or MAS.

The recent release of high‐quality chromosome‐scale assemblies and more in‐depth exploration of the THCA and CBDA biosynthesis pathway revealed a more complex modulation of their production (Grassa et al., 2021; Laverty et al., 2019; McKernan et al., 2020), primarily driven by the presence/absence, sequence variation, and expression of genes mainly located in non‐syntenic regions of chromosome 7, with additional contributions from genes on other chromosomes (Lynch et al., 2024; Van Velzen & Schranz, 2021; Weiblen et al., 2015). Therefore, it was expected that GWAS would identify the majority and most significant QTL all along chromosome 7. Nevertheless, major QTL (including SNP_3, SNP_4, SNP_20, and SNP_21) associated with THCA and CBDA were also identified on chromosomes 5, 6, and 8, such as SNP_22 on chromosome 6, which induces a missense substitution in a receptor‐like protein EIX2, with accessions carrying the beneficial alleles exhibiting twice the THC potential. This protein is typically involved in recognizing external signals and triggering plant defense responses (Sharfman et al., 2011; N. Wang et al., 2022), which correlate with the role of cannabinoids in a plant's defense system (Gorelick & Bernstein, 2017; McPartland et al., 2000; Stack et al., 2023), making LOC115696275 a promising candidate gene for THCA and CBDA quantitative modulation. In addition, SNP_7 on chromosome 2 may quantitatively influence THCA and CBDA by altering the transcriptional regulation of LOC115718628 through sequence variations in a proximal regulatory region (or cis‐regulatory element; GuhaThakurta et al., 2006; Mao et al., 2021) or by affecting posttranscriptional regulation via sequence variations in the 3′‐UTR of LOC115718629. Indeed, UTRs play a crucial role in posttranscriptional regulation by containing cis‐elements that influence various aspects of mRNA metabolism, including processing, localization, translation, and stability (Hardy & Balcerowicz, 2024). These regulatory mechanisms allow UTR variations to fine‐tune protein abundance, thereby contributing to phenotypic diversity in plants. Such variations have been associated with important agronomic traits, such as grain shape in rice (Yang et al., 2024) and fruit pigmentation in tomato (Lakshmi Jayaraj et al., 2021). Interestingly, in Cannabis, while variation in UTR sequences has been previously suggested to be associated with different chemotypes (McKernan et al., 2015), our findings indicate a potential quantitative impact, highlighting the complex role these regions may play in modulating cannabinoid production. These markers are particularly valuable for breeding programs aiming to enhance cannabinoid production. Our findings suggested that the production of THCA and CBDA is primarily influenced by codominant qualitative trait loci on chromosome 7, with additional modulation by QTLs on other chromosomes, as expected for a complex trait. For instance, the modulation of the precursor CBGA was associated with a single marker on chromosome 9, explaining a substantial impact on the phenotype within this population. Major QTLs were also identified for minor cannabinoids, including CBG, CBC, and CBN (PVE > 50%), as well as THCV and CBDV (PVE > 30%). These findings will be particularly valuable for developing varieties with customized cannabinoid profiles.

All eight significant SNPs on chromosome 7 formed a HB shared by most Type I accessions. This HB exhibited a limited recombination event over 60 Mb with six SNPs of the HB segregating in ∼95% of type I accessions. The identification of this massive haplotype was potentially influenced by ascertainment bias inherent in this GWAS panel, which is heavily skewed toward Type I accessions. This imbalance amplifies the association of the major allele haplotype with Type I, while the alternative haplotype shows a trend toward higher heterozygous form in Type II and minor alleles in Type III. Future analyses could benefit from a larger and more balanced representation of different types in the dataset, allowing for a broader evaluation of the haplotype's distribution and its potential contribution to cannabinoid modulation across different genetic backgrounds. Nevertheless, this massive haplotype aligns with previous findings of a large (∼40 Mb) non‐recombining region identified by Laverty et al. (2019) on chromosome 7 (referred to as chromosome 6 in their study), encompassing multiple repeats of THCAS and CBDAS. Similarly, massive (i.e., 1–100 Mb) non‐recombining HB associated with ecotypic differentiation were previously identified in sunflowers (Todesco et al., 2020). Unlike phenotypic plasticity, which allows for rapid acclimation to changing environmental conditions, ecotypic differentiation is a slower process that results in a genetically fixed phenotype optimized for specific environmental conditions (Franks et al., 2014; Wright et al., 2016). In the context of Cannabis cultivation, a similar phenomenon may have been stimulated by the development of phenotypically distinct Cannabis populations, that is, hemp and drug‐type, which favor distinct traits (mainly fiber vs. cannabinoids) and may thus lead to a similar massive haplotype fixation. In addition, massive HBs were often linked with large structural variants (SVs), particularly inversions, which are postulated to suppress recombination between haplotypes, thereby maintaining adaptive allelic combinations (Kirkpatrick & Barton, 2006; Todesco et al., 2020). Interestingly, in Cannabis, chromosome 7, which leads the differentiation between chemotypic profiles, hosts a massive haplotype in this GWAS panel that correlates with an enrichment in translocations and large inversions compared to other chromosomes, as demonstrated by Lynch et al. (2024). Thus, the chemotypic differentiation observed in Cannabis may be modulated by a massive haplotype on chromosome 7. In this GWAS panel, accessions carrying this massive HB associated with Type I exhibit higher THC potential, are devoid of CBDA, CBD, and CBDV, and produce less CBC compared to other accessions. This suggests that one or some SNPs within this massive haplotype on chromosome 7 are in strong LD with the causative presence/absence or sequence variation of THCAS and CBDAS genes, which are key determinants of chemotypic differentiation in Cannabis. Consequently, the massive HB on chromosome 7 could constitute a valuable tool for breeders screening germplasm for Type I accessions and for developing haplotype‐based breeding programs (Bhat et al., 2021; Rai & Tyagi, 2022).

Understanding the genetic architecture of cannabinoid biosynthesis is crucial for developing molecular breeding strategies aimed at creating new drug‐type Cannabis cultivars with specific cannabinoid profiles and optimizing cannabinoid production. Early segregation analysis of THCA and CBDA suggested that their biosynthesis was a monogenic trait modulated by codominant alleles (De Meijer et al., 2003). However, in‐depth sequencing and de novo assembly of an ever‐increasing number of Cannabis accessions have revealed a more complex genetic architecture, indicating a multigenic modulation of cannabinoid biosynthesis (Grassa et al., 2021; Laverty et al., 2019; Lynch et al., 2024; McKernan et al., 2020). Phenotypic evaluation of the GWAS panel (Lapierre et al., 2023) confirmed the qualitative production of THCA and CBDA, which was further supported by the identification of significant SNPs with qualitative impacts on these cannabinoid productions. Nevertheless, when evaluated separately, the production of THCA and CBDA can be considered quantitative traits due to the substantial variation observed among accessions of the same chemotype. In this study, the lack of signal clearly associated with canonical THCAS and CBDAS was accepted, as the GWAS panel was constituted to maximize cannabinoid and terpene diversities rather than ensure a balanced distribution of THCA and CBDA content between samples. Indeed, the limited representation of Types II and III accessions likely reduced the allele frequency of CBDAS‐associated variants, thereby diminishing the statistical power to detect significant associations. In contrast, the overrepresentation of Type I accessions may have biased the GWAS toward identifying markers associated with variation in THCA production within Type I, while limiting the ability to detect genetic determinants distinguishing THCA production in Type I from CBDA production in Type II and III. Additionally, the massive haplotype on chromosome 7, which spans a broad region including several THCAS and CBDAS, may obscure fine‐scale resolution, particularly in the likely presence of large SVs such as inversions and translocations. Furthermore, given the nature of the GWAS methodology, GWAS alone is not designed to identify causal genes, especially for complex traits influenced by polygenic effects. Thus, it could be expected that key enzymes involved in the cannabinoid biosynthetic pathway were not identified in the vicinity of the significant SNPs. Expanding the population and incorporating complementary approaches, such as haplotype‐based analysis and transcriptomics, could enhance the resolution of candidate gene identification, potentially improving the detection of THCAS and CBDAS.

Complex traits could be challenging for GWAS due to their intricate nature, influenced by numerous loci, potential epistasis, and gene‐environment interactions (Gupta et al., 2014; Zhou & Huang, 2019). These complexities can obscure significant associations, necessitating advanced statistical models to accurately capture the genetic architecture of these traits. The BLINK model, recognized as one of the most statistically powerful approaches for multi‐locus GWAS in plants (Kaler et al., 2020; J. Wang & Zhang, 2021), is particularly suited for conducting GWAS on cannabinoid traits due to its ability to capture complex interactions among multiple loci (Huang et al., 2019). While large HBs could affect the genome‐wide estimates of relatedness between accessions, potentially masking association signals in GWAS (Lotterhos, 2019), the BLINK model has proven effective in detecting several markers with high significance, particularly along chromosome 7, which hosts a massive HB.

The substantial variation in heterozygous genotypes between accessions suggested differing degrees of inbreeding and/or outcrossing within the GWAS population, independent of population structure or chemotypes. The relatively uniform nucleotide diversity observed across autosomes, while chromosome X exhibits the lowest diversity, was consistent with previous findings of significantly slower LD decay on chromosome X compared to autosomes (p < 0.001), likely reflecting breeding practices that emphasize sexual traits, such as female flower production (de Ronne et al., 2024). In contrast, the absence of expected diversity reduction on chromosome 7, where THCAS and CBDAS are located, should be nuanced here, as the GWAS panel was highly imbalanced between chemotypes, which may have biased diversity estimates.

Finally, using a reference genome for GWAS allows the identification of molecular markers at a precise location in the genome and therefore facilitates comparison with other marker‐association studies and cross‐validation through meta‐analysis of GWAS summary statistics (MetaGWAS; Shook et al., 2021). These markers are promising candidates for breeding programs using MAS, which is expected to accelerate the development of new Cannabis varieties with enhanced and specific cannabinoid profiles tailored for medical and recreational uses. Further functional validation of putative causal SNPs and candidate genes could contribute to improve our understanding of cannabinoid biosynthesis in Cannabis.

5. CONCLUSION

To address the shortage of genetic markers associated with cannabinoid traits, which are essential for Cannabis research and industrial applications, this study provided a comprehensive catalog of SNPs that have major impacts on cannabinoid biosynthesis. Additionally, a massive haplotype discriminating Type I accession was identified on chromosome 7. These molecular markers will constitute an essential tool in breeding programs that utilize MAS. They promise to accelerate the selection process for promising accessions, potential crossing parents, while significantly reducing costs associated with labor‐intensive phenotype‐based selection methods. This advancement not only enhances our understanding of Cannabis genetics but also streamlines the development of cultivars with desired cannabinoid profiles, thereby supporting the growing needs of medical, recreational, and industrial sectors.

AUTHOR CONTRIBUTIONS

Maxime de Ronne: Conceptualization; data curation; formal analysis; investigation; methodology; visualization; writing—original draft; writing—review and editing. Davoud Torkamaneh: Conceptualization; funding acquisition; project administration; supervision; visualization; writing—review and editing.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Table S1: Description of cannabinoid traits in the 174 drug‐type Cannabis accessions. Data were provided by Lapierre et al. (2023). THCV, CBDV and CBN are qualitative variables, with 0 indicating absence and 1 indicating presence.

Table S2: Genome‐ and chromosome‐wide nucleotide variation analysis across the GWAS panel, clustering assignments and chemotypes. The 154 Cannabis accessions were distributed as K1 = 35, K2 = 52 and K3 = 87, Type I = 152, Type II = 17 and Type III = 5.

Table S3: Identification of candidate genes through haplotype block (HB) analysis and functional impact analysis of SNPs. Orthology analysis of Cannabis genes was performed against the Arabidopsis thaliana transcriptome TAIR 11 (Cheng et al., 2017). Only significant SNPs and SNP residing in HB of interest located in a gene sequence were considered for functional impact analysis (detailed in Supplemental Table 3).

Table S4: Functional variant analysis of significant SNPs and SNP residing in HB of interest located in a gene sequence. The SNPs having a putative impact on gene function were bolded.

Figure S1: Heatmap of pairwise kinship matrix values between the 174 drug‐type accessions. The color histogram showsthe distribution of co‐ancestry coefficients. The redder the color, the more similar the individuals are, whereas the more yellow the color, the more dissimilar the individuals are.

Figure S2: Discriminant analysis of principal components (DAPC) of the 174 drug‐type accessions using the whole 282K high‐quality SNPs. a. Cumulated variance explained by the eigenvalues of the PCA. b. Value of BIC versus for increasing values of k. Cluster selection was based on the BIC value. c. Optimization α‐score graph. The optimal α‐score suggested that twenty PCs should be retained for the discriminant analysis. d. DAPC cross‐validation test for the optimal number of PCs retained.

Figure S3: Population structure analysis of the 174 drug‐type accessions using 282K high‐quality SNPs. Dark blue, light blue and green correspond to cluster K1, K2 and K3, respectively. a) Admixture plot for K = 3 using fastStructure. Each vertical line represent an accession. b) Discriminant analysis of principal components (DAPC) scatter plot. c) Unrooted phylogenetic tree estimated using the Unweighted Pair Group Method with Arithmetic Mean (UPGMA). Colors are based on fastStructure clustering. d) Concordance of cluster assignment for fastStructure and DAPC inference.

Figure S4: Quantil‐quantil (QQ) plots for eleven cannabinoid traits in drug‐type Cannabis.

Figure S5: Distribution of allelic classes of significant markers associated with cannabinoid traits in 174 drug‐type Cannabis accessions.

Figure S6: Distribution of allelic classes of 5 SNPs with putative impact on gene function.

Figure S7: Distribution of allelic classes of significant SNPs compared to chemotype categories. Chemotype are based on the THC/CBDratio. Major alleles are highlighted in blue, minor alleles in orange, and heterozygous alleles in grey.

TPG2-18-e70031-s001.xlsx (41.4KB, xlsx)

ACKNOWLEDGMENTS

The authors wish to thank Justine Richard‐Giroux, Éliana Lapierre, and Perrine Feutry for their valuable contributions to the tedious phenotyping and data collection. This work was conducted as part of a collaborative research project funded by Fuga Group Inc. and NSERC Alliance [#ALLRP 568653–21 to D.T.].

de Ronne, M. , & Torkamaneh, D. (2025). Discovery of major QTL and a massive haplotype associated with cannabinoid biosynthesis in drug‐type Cannabis . The Plant Genome, 18, e70031. 10.1002/tpg2.70031

Assigned to Associate Editor Nils Stein.

DATA AVAILABILITY STATEMENT

The sequencing files generated used for the analyzes of this study are on NCBI BioProject: PRJNA1047573 and will be accessible after acceptance of the manuscript (reviewer link: https://dataview.ncbi.nlm.nih.gov/object/PRJNA1047573?reviewer=kr7i954sp11ttcn4qm36trajc4).

REFERENCES

  1. Aboul‐Maaty, N. A.‐F. , & Oraby, H. A.‐S. (2019). Extraction of high‐quality genomic DNA from different plant orders applying a modified CTAB‐based method. Bulletin of the National Research Centre, 43(1), Article 25. 10.1186/S42269-019-0066-1 [DOI] [Google Scholar]
  2. Bai, Y. , Jiang, M. , Xie, T. , Jiang, C. , Gu, M. , Zhou, X. , Yan, X. , Yuan, Y. , & Huang, L. (2021). Archaeobotanical evidence of the use of medicinal cannabis in a secular context unearthed from south China. Journal of ethnopharmacology, 275, 114114. 10.1016/J.JEP.2021.114114 [DOI] [PubMed] [Google Scholar]
  3. Bakker, E. , Holloway, A. , & K Waterman—US Patent App. 17/665, 500, & 2023, U . (2021). Autoflowering markers (US12024712B2). Google Patents. https://patents.google.com/patent/US12024712B2/en [Google Scholar]
  4. Baratta, F. , Pignata, I. , Ravetto Enri, L. , & Brusa, P. (2022). Cannabis for medical use: Analysis of recent clinical trials in view of current legislation. Frontiers in Pharmacology, 13, 888903. 10.3389/FPHAR.2022.888903 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Benjamini, Y. , & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B (Methodological), 57, 289–300. 10.1111/J.2517-6161.1995.TB02031.X [DOI] [Google Scholar]
  6. Bewley‐Taylor, D. , & Jelsma, M. (2011). Regime change: Re‐visiting the 1961 single convention on narcotic drugs. International Journal of Drug Policy, 23, 72–81. 10.1016/j.drugpo.2011.08.003 [DOI] [PubMed] [Google Scholar]
  7. Bhat, J. A. , Yu, D. , Bohra, A. , Ganie, S. A. , & Varshney, R. K. (2021). Features and applications of haplotypes in crop breeding. Communications Biology, 4(1), Article 1266. 10.1038/s42003-021-02782-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bradbury, P. J. , Zhang, Z. , Kroon, D. E. , Casstevens, T. M. , Ramdoss, Y. , & Buckler, E. S. (2007). TASSEL: Software for association mapping of complex traits in diverse samples. Bioinformatics, 23, 2633–2635. 10.1093/bioinformatics/btm308 [DOI] [PubMed] [Google Scholar]
  9. Browning, B. L. , & Browning, S. R. (2016). Genotype imputation with millions of reference samples. American Journal of Human Genetics, 98, 116–126. 10.1016/J.AJHG.2015.11.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Carey, S. B. , Jenkins, J. , Lovell, J. T. , Maumus, F. , Sreedasyam, A. , Payton, A. C. , Shu, S. , Tiley, G. P. , Fernandez‐Pozo, N. , Healey, A. , Barry, K. , Chen, C. , Wang, M. , Lipzen, A. , Daum, C. , Saski, C. A. , McBreen, J. C. , Conrad, R. E. , Kollar, L. M. , … McDaniel, S. F. (2021). Gene‐rich UV sex chromosomes harbor conserved regulators of sexual development. Science Advances, 7, eabh2488. 10.1126/SCIADV.ABH2488/SUPPL_FILE/ABH2488_TABLES_S1_TO_S17.XLSX [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cheng, C. Y. , Krishnakumar, V. , Chan, A. P. , Thibaud‐Nissen, F. , Schobel, S. , & Town, C. D. (2017). Araport11: A complete reannotation of the Arabidopsis thaliana reference genome. The Plant Journal, 89, 789–804. 10.1111/TPJ.13415 [DOI] [PubMed] [Google Scholar]
  12. Clarke, R. , & Merlin, M. (2016). Cannabis: Evolution and ethnobotany. University of California Press. [Google Scholar]
  13. Coelho, M. P. , Duarte, P. , Calado, M. , Almeida, A. J. , Reis, C. P. , & Gaspar, M. M. (2023). The current role of cannabis and cannabinoids in health: A comprehensive review of their therapeutic potential. Life Sciences, 329, 121838. 10.1016/J.LFS.2023.121838 [DOI] [PubMed] [Google Scholar]
  14. Collard, B. C. Y. , & Mackill, D. J. (2008). Marker‐assisted selection: An approach for precision plant breeding in the twenty‐first century. Philosophical Transactions of the Royal Society B: Biological Sciences, 363, 557–572. 10.1098/rstb.2007.2170 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Competition Bureau Canada . (2023). Impact on the Canadian economy. In Planting the seeds for competition (p. 14). ISED Citizen Services Centre. https://ised‐isde.canada.ca/site/competition‐bureau‐canada/en/how‐we‐foster‐competition/education‐and‐outreach/planting‐seeds‐competition [Google Scholar]
  16. Cox, C. (2018). The Canadian Cannabis Act legalizes and regulates recreational cannabis use in 2018. Health Policy, 122, 205–209. 10.1016/J.HEALTHPOL.2018.01.009 [DOI] [PubMed] [Google Scholar]
  17. Danecek, P. , Auton, A. , Abecasis, G. , Albers, C. A. , Banks, E. , DePristo, M. A. , Handsaker, R. E. , Lunter, G. , Marth, G. T. , Sherry, S. T. , McVean, G. , & Durbin, R. (2011). The variant call format and VCFtools. Bioinformatics, 27, 2156–2158. 10.1093/bioinformatics/btr330 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. De La Vega, F. M. , Lazaruk, K. D. , Rhodes, M. D. , & Wenz, M. H. (2005). Assessment of two flexible and compatible SNP genotyping platforms: TaqMan® SNP genotyping assays and the SNPlex™ genotyping system. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis, 573, 111–135. 10.1016/J.MRFMMM.2005.01.008 [DOI] [PubMed] [Google Scholar]
  19. de Koning, D. J. (2016). Meuwissen et al. On genomic selection. Genetics, 203, 5–7. 10.1534/genetics.116.189795 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. De Meijer, E. P. M. , Bagatta, M. , Carboni, A. , Crucitti, P. , Moliterni, V. M. C. , Ranalli, P. , & Mandolino, G. (2003). The inheritance of chemical phenotype in Cannabis sativa L. Genetics, 163, 335–346. 10.1093/GENETICS/163.1.335 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. de Ronne, M. , Lapierre, É. , & Torkamaneh, D. (2024). Genetic insights into agronomic and morphological traits of drug‐type cannabis revealed by genome‐wide association studies. Scientific Reports, 14, Article 9162. 10.1038/S41598-024-58931-W [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Eilbeck, K. , Lewis, S. E. , Mungall, C. J. , Yandell, M. , Stein, L. , Durbin, R. , & Ashburner, M. (2005). The sequence ontology: A tool for the unification of genome annotations. Genome Biology, 6, Article R44. 10.1186/GB-2005-6-5-R44/FIGURES/4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Emms, D. M. , & Kelly, S. (2019). OrthoFinder: Phylogenetic orthology inference for comparative genomics. Genome Biology, 20, Article 238. 10.1186/S13059-019-1832-Y/FIGURES/5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Fellermeier, M. , Eisenreich, W. , Bacher, A. , & Zenk, M. H. (2001). Biosynthesis of cannabinoids. Incorporation experiments with (13)C‐labeled glucoses. European Journal of Biochemistry, 268, 1596–1604. 10.1046/J.1432-1033.2001.02030.X [DOI] [PubMed] [Google Scholar]
  25. Francia, E. , Tacconi, G. , Crosatti, C. , Barabaschi, D. , Bulgarelli, D. , Dall'Aglio, E. , & Valè, G. (2005). Marker assisted selection in crop plants. Plant Cell, Tissue and Organ Culture, 82, 317–342. 10.1007/s11240-005-2387-z [DOI] [Google Scholar]
  26. Franks, S. J. , Weber, J. J. , & Aitken, S. N. (2014). Evolutionary and plastic responses to climate change in terrestrial plant populations. Evolutionary Applications, 7, 123–139. 10.1111/EVA.12112 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Gagne, S. J. , Stout, J. M. , Liu, E. , Boubakir, Z. , Clark, S. M. , & Page, J. E. (2012). Identification of olivetolic acid cyclase from Cannabis sativa reveals a unique catalytic route to plant polyketides. Proceedings of the National Academy of Sciences of the United States of America, 109, 12811–12816. 10.1073/PNAS.1200330109 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Geissler, M. , Volk, J. , Stehle, F. , Kayser, O. , & Warzecha, H. (2018). Subcellular localization defines modification and production of Δ9‐tetrahydrocannabinolic acid synthase in transiently transformed Nicotiana benthamiana . Biotechnology Letters, 40, 981–987. 10.1007/S10529-018-2545-0 [DOI] [PubMed] [Google Scholar]
  29. Gorelick, J. , & Bernstein, N. (2017). Chemical and physical elicitation for enhanced cannabinoid production in Cannabis . In Chandra S., Lata H., & ElSohly M. A. (Eds.), Cannabis sativa L.—Botany and biotechnology (pp. 439–456). Springer. 10.1007/978-3-319-54564-6_21 [DOI] [Google Scholar]
  30. Grassa, C. J. , Weiblen, G. D. , Wenger, J. P. , Dabney, C. , Poplawski, S. G. , Timothy Motley, S. , Michael, T. P. , & Schwartz, C. J. (2021). A new Cannabis genome assembly associates elevated cannabidiol (CBD) with hemp introgressed into marijuana. New Phytologist, 230, 1665–1679. 10.1111/NPH.17243 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Grassa, C. J. , Wenger, J. P. , Dabney, C. , Poplawski, S. G. , Motley, S. T. , Michael, T. P. , Schwartz, C. J. , & Weiblen, G. D. (2018). A complete Cannabis chromosome assembly and adaptive admixture for elevated cannabidiol (CBD) content. bioRxiv. 10.1101/458083 [DOI]
  32. GuhaThakurta, D. , Xie, T. , Anand, M. , Edwards, S. W. , Li, G. , Wang, S. S. , & Schadt, E. E. (2006). Cis‐regulatory variations: A study of SNPs around genes showing cis‐linkage in segregating mouse populations. BMC Genomics, 7, Article 235. 10.1186/1471-2164-7-235 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Gülck, T. , & Møller, B. L. (2020). Phytocannabinoids: Origins and biosynthesis. Trends in Plant Science, 25, 985–1004. 10.1016/J.TPLANTS.2020.05.005 [DOI] [PubMed] [Google Scholar]
  34. Gupta, P. K. , Kulwal, P. L. , & Jaiswal, V. (2014). Association mapping in crop plants: Opportunities and challenges. Advances in Genetics, 85, 109–147. 10.1016/B978-0-12-800271-1.00002-0 [DOI] [PubMed] [Google Scholar]
  35. Hanuš, L. O. , Meyer, S. M. , Muñoz, E. , Taglialatela‐Scafati, O. , & Appendino, G. (2016). Phytocannabinoids: A unified critical inventory. Natural Product Reports, 33, 1357–1392. 10.1039/C6NP00074F [DOI] [PubMed] [Google Scholar]
  36. Hardy, E. C. , & Balcerowicz, M. (2024). Untranslated yet indispensable—UTRs act as key regulators in the environmental control of gene expression. Journal of Experimental Botany, 75, 4314–4331. 10.1093/JXB/ERAE073 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Hesami, M. , Pepe, M. , Alizadeh, M. , Rakei, A. , Baiton, A. , & Phineas Jones, A. M. (2020). Recent advances in cannabis biotechnology. Industrial Crops and Products, 158, 113026. 10.1016/j.indcrop.2020.113026 [DOI] [Google Scholar]
  38. Huang, M. , Liu, X. , Zhou, Y. , Summers, R. M. , & Zhang, Z. (2019). BLINK: A package for the next level of genome‐wide association studies with both individuals and markers in the millions. GigaScience, 8, giy154. 10.1093/gigascience/giy154 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Hurgobin, B. , Tamiru‐Oli, M. , Welling, M. T. , Doblin, M. S. , Bacic, A. , Whelan, J. , & Lewsey, M. G. (2021). Recent advances in Cannabis sativa genomics research. New Phytologist, 230, 73–89. 10.1111/NPH.17140 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Jannink, J. L. , Lorenz, A. J. , & Iwata, H. (2010). Genomic selection in plant breeding: From theory to practice. Briefings in Functional Genomics and Proteomics, 9, 166–177. 10.1093/bfgp/elq001 [DOI] [PubMed] [Google Scholar]
  41. Jombart, T. , Devillard, S. , & Balloux, F. (2010). Discriminant analysis of principal components: A new method for the analysis of genetically structured populations. BMC Genetics, 11, Article 94. 10.1186/1471-2156-11-94/FIGURES/9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Kaler, A. S. , Gillman, J. D. , Beissinger, T. , & Purcell, L. C. (2020). Comparing different statistical models and multiple testing corrections for association mapping in soybean and maize. Frontiers in Plant Science, 10, Article 1794. 10.3389/fpls.2019.01794 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Kim, A. L. , Yun, Y. J. , Choi, H. W. , Hong, C. H. , Shim, H. J. , Lee, J. H. , & Kim, Y. C. (2022). Profiling cannabinoid contents and expression levels of corresponding biosynthetic genes in commercial cannabis (Cannabis sativa L.) cultivars. Plants, 11, 3088. 10.3390/PLANTS11223088/S1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Kirkpatrick, M. , & Barton, N. (2006). Chromosome inversions, local adaptation and speciation. Genetics, 173, 419–434. 10.1534/GENETICS.105.047985 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Kovalchuk, I. , Pellino, M. , Rigault, P. , Van Velzen, R. , Ebersbach, J. , Ashnest, J. R. , Mau, M. , Schranz, M. E. , Alcorn, J. , Laprairie, R. B. , McKay, J. K. , Burbridge, C. , Schneider, D. , Vergara, D. , Kane, N. C. , & Sharbel, T. F. (2020). The genomics of Cannabis and its close relatives. Annual Review of Plant Biology, 71, 713–739. 10.1146/ANNUREV-ARPLANT-081519-040203 [DOI] [PubMed] [Google Scholar]
  46. Lakshmi Jayaraj, K. , Thulasidharan, N. , Antony, A. , John, M. , Augustine, R. , Chakravartty, N. , Sukumaran, S. , Uma Maheswari, M. , Abraham, S. , Thomas, G. , Lachagari, V. B. R. , Seshagiri, S. , Narayanan, S. , & Kuriakose, B. (2021). Targeted editing of tomato carotenoid isomerase reveals the role of 5′ UTR region in gene expression regulation. Plant Cell Reports, 40, 621–635. 10.1007/S00299-020-02659-0/METRICS [DOI] [PubMed] [Google Scholar]
  47. Lapierre, É. , de Ronne, M. , Boulanger, R. , & Torkamaneh, D. (2023). Comprehensive phenotypic characterization of diverse drug‐type Cannabis varieties from the Canadian legal market. Plants, 12, 3756. 10.3390/PLANTS12213756 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Laverty, K. U. , Stout, J. M. , Sullivan, M. J. , Shah, H. , Gill, N. , Holbrook, L. , Deikus, G. , Sebra, R. , Hughes, T. R. , Page, J. E. , & Van Bakel, H. (2019). A physical and genetic map of Cannabis sativa identifies extensive rearrangements at the THC/CBD acid synthase loci. Genome Research, 29, 146–156. 10.1101/GR.242594.118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Lotterhos, K. E. (2019). The effect of neutral recombination variation on genome scans for selection. G3 Genes|Genomes|Genetics, 9, 1851–1867. 10.1534/G3.119.400088 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Lu, H. C. , & MacKie, K. (2016). An introduction to the endogenous cannabinoid system. Biological Psychiatry, 79, 516–525. 10.1016/J.BIOPSYCH.2015.07.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Luo, X. , Reiter, M. A. , d'Espaux, L. , Wong, J. , Denby, C. M. , Lechner, A. , Zhang, Y. , Grzybowski, A. T. , Harth, S. , Lin, W. , Lee, H. , Yu, C. , Shin, J. , Deng, K. , Benites, V. T. , Wang, G. , Baidoo, E. E. K. , Chen, Y. , Dev, I. , … Keasling, J. D. (2019). Complete biosynthesis of cannabinoids and their unnatural analogues in yeast. Nature, 567(7746), 123–126. 10.1038/s41586-019-0978-9 [DOI] [PubMed] [Google Scholar]
  52. Lynch, R. C. , Padgitt‐Cobb, L. K. , Garfinkel, A. R. , Knaus, B. J. , Hartwick, N. T. , Allsing, N. , Aylward, A. , Mamerto, A. , Kitony, J. K. , Colt, K. , Murray, E. R. , Duong, T. , Trippe, A. , Crawford, S. , Vining, K. , Michael, T. P. , & Andrea Garfinkel, O. R. (2024). Domesticated cannabinoid synthases amid a wild mosaic Cannabis pangenome. bioRxiv. 10.1101/2024.05.21.595196 [DOI] [Google Scholar]
  53. Mao, H. , Li, S. , Chen, B. , Jian, C. , Mei, F. , Zhang, Y. , Li, F. , Chen, N. , Li, T. , Du, L. , Ding, L. , Wang, Z. , Cheng, X. , Wang, X. , & Kang, Z. (2021). Variation in cis‐regulation of a NAC transcription factor contributes to drought tolerance in wheat. Molecular Plant, 15, 276–292. 10.1016/J.MOLP.2021.11.007 [DOI] [PubMed] [Google Scholar]
  54. Martínez, V. , Iriondo De‐Hond, A. , Borrelli, F. , Capasso, R. , Del Castillo, M. D. , & Abalo, R. (2020). Cannabidiol and other non‐psychoactive cannabinoids for prevention and treatment of gastrointestinal disorders: Useful nutraceuticals? International Journal of Molecular Sciences, 21, 3067. 10.3390/IJMS21093067 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. McKernan, K. J. , Helbert, Y. , Kane, L. T. , Ebling, H. , Zhang, L. , Liu, B. , Eaton, Z. , McLaughlin, S. , Kingan, S. , Baybayan, P. , Concepcion, G. , Jordan, M. , Riva, A. , Barbazuk, W. , & Harkins, T. (2020). Sequence and annotation of 42 Cannabis genomes reveals extensive copy number variation in cannabinoid synthesis and pathogen resistance genes . bioRxiv. 10.1101/2020.01.03.894428 [DOI]
  56. McKernan, K. J. , Helbert, Y. , Tadigotla, V. , McLaughlin, S. , Spangler, J. , Zhang, L. , & Smith, D. (2015). Single molecule sequencing of THCA synthase reveals copy number variation in modern drug‐type Cannabis sativa L. bioRxiv. 10.1101/028654 [DOI]
  57. McLaren, W. , Gil, L. , Hunt, S. E. , Riat, H. S. , Ritchie, G. R. S. , Thormann, A. , Flicek, P. , & Cunningham, F. (2016). The Ensembl variant effect predictor. Genome Biology, 17, Article 122. 10.1186/s13059-016-0974-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. McPartland, J. M. , Clarke, R. C. , & Watson, D. P. (2000). Hemp diseases and pests. Integrated Environmental Assessment and Management, 11, 276. [Google Scholar]
  59. Mechoulam, R. , & Parker, L. A. (2013). The endocannabinoid system and the brain. Annual Review of Psychology, 64, 21–47. 10.1146/ANNUREV-PSYCH-113011-143739 [DOI] [PubMed] [Google Scholar]
  60. Monthony, A. S. , de Ronne, M. , & Torkamaneh, D. (2024). Exploring ethylene‐related genes in Cannabis sativa: Implications for sexual plasticity. Plant Reproduction, 37(3), 321–339. 10.1007/s00497-023-00492-5 [DOI] [PubMed] [Google Scholar]
  61. Morimoto, S. , Komatsu, K. , Taura, F. , & Shoyama, Y. (1997). Enzymological evidence for cannabichromenic acid biosynthesis. Journal of Natural Products, 60, 854–867. 10.1021/NP970210Y [DOI] [Google Scholar]
  62. Morrell, P. L. , Buckler, E. S. , & Ross‐Ibarra, J. (2012). Crop genomics: Advances and applications. Nature Reviews Genetics, 13, 85–96. 10.1038/nrg3097 [DOI] [PubMed] [Google Scholar]
  63. NCBI . (2019). NCBI Cannabis sativa annotation release 100 . NCBI. https://www.ncbi.nlm.nih.gov/genome/annotation_euk/Cannabis_sativa/100/ [Google Scholar]
  64. Nei, M. , & Li, W. H. (1979). Mathematical model for studying genetic variation in terms of restriction endonucleases. Proceedings of the National Academy of Sciences, 76, 5269–5273. 10.1073/PNAS.76.10.5269 [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Pertwee, R. G. (2008). The diverse CB1 and CB2 receptor pharmacology of three plant cannabinoids: Δ9‐tetrahydrocannabinol, cannabidiol and Δ9‐tetrahydrocannabivarin. British Journal of Pharmacology, 153, 199–215. 10.1038/SJ.BJP.0707442 [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Petit, J. , Salentijn, E. M. J. , Paulo, M. J. , Denneboom, C. , & Trindade, L. M. (2020). Genetic architecture of flowering time and sex determination in hemp (Cannabis sativa L.): A genome‐wide association study. Frontiers in Plant Science, 11, 569958. 10.3389/FPLS.2020.569958/BIBTEX [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Petit, J. , Salentijn, E. M. J. , Paulo, M. J. , Denneboom, C. , van Loo, E. N. , & Trindade, L. M. (2020). Elucidating the genetic architecture of fiber quality in hemp (Cannabis sativa L.) using a genome‐wide association study. Frontiers in Genetics, 11, 566314. 10.3389/FGENE.2020.566314 [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Purcell, S. , Neale, B. , Todd‐Brown, K. , Thomas, L. , Ferreira, M. A. R. , Bender, D. , Maller, J. , Sklar, P. , De Bakker, P. I. W. , Daly, M. J. , & Sham, P. C. (2007). PLINK: A tool set for whole‐genome association and population‐based linkage analyses. American Journal of Human Genetics, 81, 559–575. 10.1086/519795 [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. R Core Team . (2021). R: The R project for statistical computing . https://www.r‐project.org/
  70. Rai, M. , & Tyagi, W. (2022). Haplotype breeding for unlocking and utilizing plant genomics data. Frontiers in Genetics, 13, 1006288. 10.3389/FGENE.2022.1006288 [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Raj, A. , Stephens, M. , & Pritchard, J. K. (2014). FastSTRUCTURE: Variational inference of population structure in large SNP data sets. Genetics, 197, 573–589. 10.1534/genetics.114.164350 [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Rock, E. M. , & Parker, L. A. (2021). Constituents of Cannabis Sativa . In Murillo‐Rodriguez E., Pandi‐Perumal S. R., & Monti J. M. (Eds.), Cannabinoids and neuropsychiatric disorders (Vol. 1264, pp. 1–13). Advances in experimental medicine and biology. Springer. 10.1007/978-3-030-57369-0_1 [DOI] [PubMed] [Google Scholar]
  73. Rodziewicz, P. , Loroch, S. , Marczak, Ł. , Sickmann, A. , & Kayser, O. (2019). Cannabinoid synthases and osmoprotective metabolites accumulate in the exudates of Cannabis sativa L. glandular trichomes. Plant Science, 284, 108–116. 10.1016/J.PLANTSCI.2019.04.008 [DOI] [PubMed] [Google Scholar]
  74. Russo, E. B. (2016). Beyond Cannabis: Plants and the endocannabinoid system. Trends in Pharmacological Sciences, 37, 594–605. 10.1016/J.TIPS.2016.04.005 [DOI] [PubMed] [Google Scholar]
  75. Semagn, K. , Babu, R. , Hearne, S. , & Olsen, M. (2014). Single nucleotide polymorphism genotyping using kompetitive allele specific PCR (KASP): Overview of the technology and its application in crop improvement. Molecular Breeding, 33, 1–14. 10.1007/s11032-013-9917-x [DOI] [Google Scholar]
  76. Sharfman, M. , Bar, M. , Ehrlich, M. , Schuster, S. , Melech‐Bonfil, S. , Ezer, R. , Sessa, G. , & Avni, A. (2011). Endosomal signaling of the tomato leucine‐rich repeat receptor‐like protein LeEix2. The Plant Journal, 68, 413–423. 10.1111/J.1365-313X.2011.04696.X [DOI] [PubMed] [Google Scholar]
  77. Shook, J. M. , Zhang, J. , Jones, S. E. , Singh, A. , Diers, B. W. , & Singh, A. K. (2021). Meta‐GWAS for quantitative trait loci identification in soybean. G3 Genes|Genomes|Genetics, 11, jkab117. 10.1093/G3JOURNAL/JKAB117 [DOI] [PMC free article] [PubMed] [Google Scholar]
  78. Sirikantaramas, S. , Morimoto, S. , Shoyama, Y. , Ishikawa, Y. , Wada, Y. , Shoyama, Y. , & Taura, F. (2004). The gene controlling marijuana psychoactivity. Molecular cloning and heterologous expression of Δ1‐tetrahydrocannabinolic acid synthase from Cannabis sativa L.. Journal of Biological Chemistry, 279, 39767–39774. 10.1074/jbc.M403693200 [DOI] [PubMed] [Google Scholar]
  79. Sirikantaramas, S. , Taura, F. , Tanaka, Y. , Ishikawa, Y. , Morimoto, S. , & Shoyama, Y. (2005). Tetrahydrocannabinolic acid synthase, the enzyme controlling marijuana psychoactivity, is secreted into the storage cavity of the glandular trichomes. Plant and Cell Physiology, 46, 1578–1582. 10.1093/PCP/PCI166 [DOI] [PubMed] [Google Scholar]
  80. Stack, G. M. , Snyder, S. I. , Toth, J. A. , Quade, M. A. , Crawford, J. L. , McKay, J. K. , Jackowetz, J. N. , Wang, P. , Philippe, G. , Hansen, J. L. , Moore, V. M. , Rose, J. K. C. , & Smart, L. B. (2023). Cannabinoids function in defense against chewing herbivores in Cannabis sativa L. Horticulture Research, 10, uhad207. 10.1093/HR/UHAD207 [DOI] [PMC free article] [PubMed] [Google Scholar]
  81. Storr, M. A. , & Sharkey, K. A. (2007). The endocannabinoid system and gut–brain signalling. Current Opinion in Pharmacology, 7, 575–582. 10.1016/J.COPH.2007.08.008 [DOI] [PubMed] [Google Scholar]
  82. Sun, J. , Chen, J. , Zhang, X. , Xu, G. , Yu, Y. , Dai, Z. , & Su, J. (2023). Genome‐wide association study of salt tolerance at the germination stage in hemp. Euphytica, 219, Article 5. 10.1007/s10681-022-03129-2 [DOI] [Google Scholar]
  83. Tan, Z. , Clomburg, J. M. , & Gonzalez, R. (2018). Synthetic pathway for the production of olivetolic acid in Escherichia coli . ACS Synthetic Biology, 7, 1886–1896. 10.1021/acssynbio.8b00075 [DOI] [PubMed] [Google Scholar]
  84. Taura, F. , Sirikantaramas, S. , Shoyama, Y. , Yoshikai, K. , Shoyama, Y. , & Morimoto, S. (2007). Cannabidiolic‐acid synthase, the chemotype‐determining enzyme in the fiber‐type Cannabis sativa . FEBS letters, 581, 2929–2934. 10.1016/J.FEBSLET.2007.05.043 [DOI] [PubMed] [Google Scholar]
  85. Todesco, M. , Owens, G. L. , Bercovich, N. , Légaré, J. S. , Soudi, S. , Burge, D. O. , Huang, K. , Ostevik, K. L. , Drummond, E. B. M. , Imerovski, I. , Lande, K. , Pascual‐Robles, M. A. , Nanavati, M. , Jahani, M. , Cheung, W. , Staton, S. E. , Muños, S. , Nielsen, R. , Donovan, L. A. , … Rieseberg, L. H. (2020). Massive haplotypes underlie ecotypic differentiation in sunflowers. Nature, 584(7822), 602–607. 10.1038/s41586-020-2467-6 [DOI] [PubMed] [Google Scholar]
  86. Torkamaneh, D. , & Jones, A. M. P. (2021). Cannabis, the multibillion dollar plant that no genebank wanted. Genome, 65, 1–5. 10.1139/gen-2021-0016 [DOI] [PubMed] [Google Scholar]
  87. Torkamaneh, D. , Laroche, J. , & Belzile, F. (2020). Fast‐gbs v2.0: An analysis toolkit for genotyping‐by‐sequencing data. Genome, 63, 577–581. 10.1139/gen-2020-0077 [DOI] [PubMed] [Google Scholar]
  88. Torkamaneh, D. , Laroche, J. , Boyle, B. , Hyten, D. L. , & Belzile, F. (2021). A bumper crop of SNPs in soybean through high‐density genotyping‐by‐sequencing (HD‐GBS). Plant Biotechnology Journal, 19, 860–862. 10.1111/pbi.13551 [DOI] [PMC free article] [PubMed] [Google Scholar]
  89. Van Velzen, R. , & Schranz, M. E. (2021). Origin and evolution of the Cannabinoid Oxidocyclase gene family. Genome Biology and Evolution, 13, evab130. 10.1093/GBE/EVAB130 [DOI] [PMC free article] [PubMed] [Google Scholar]
  90. Wang, J. , & Zhang, Z. (2021). GAPIT version 3: Boosting power and accuracy for genomic association and prediction. Genomics, Proteomics & Bioinformatics, 19, 629–640. 10.1016/J.GPB.2021.08.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  91. Wang, N. , Yin, Z. , Zhao, Y. , Li, Z. , Dou, D. , & Wei, L. (2022). Two divergent immune receptors of the allopolyploid Nicotiana benthamiana reinforce the recognition of a fungal microbe‐associated molecular pattern VdEIX3. Frontiers in Plant Science, 13, 968562. 10.3389/FPLS.2022.968562/BIBTEX [DOI] [PMC free article] [PubMed] [Google Scholar]
  92. Watts, S. , McElroy, M. , Migicovsky, Z. , Maassen, H. , van Velzen, R. , & Myles, S. (2021). Cannabis labelling is associated with genetic variation in terpene synthase genes. Nature Plants, 7(10), 1330–1334. 10.1038/s41477-021-01003-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  93. Weiblen, G. D. , Wenger, J. P. , Craft, K. J. , ElSohly, M. A. , Mehmedic, Z. , Treiber, E. L. , & Marks, M. D. (2015). Gene duplication and divergence affecting drug content in Cannabis sativa . The New phytologist, 208, 1241–1250. 10.1111/NPH.13562 [DOI] [PubMed] [Google Scholar]
  94. Welling, M. T. , Liu, L. , Kretzschmar, T. , Mauleon, R. , Ansari, O. , & King, G. J. (2020). An extreme‐phenotype genome‐wide association study identifies candidate cannabinoid pathway genes in Cannabis . Scientific Reports, 10(1), Article 18643. 10.1038/s41598-020-75271-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  95. Welling, M. T. , Shapter, T. , Rose, T. J. , Liu, L. , Stanger, R. , & King, G. J. (2016). A belated green revolution for Cannabis: Virtual genetic resources to fast‐track cultivar development. Frontiers in Plant Science, 7, 205761. 10.3389/FPLS.2016.01113/BIBTEX [DOI] [PMC free article] [PubMed] [Google Scholar]
  96. Wright, J. P. , Ames, G. M. , & Mitchell, R. M. (2016). The more things change, the more they stay the same? When is trait variability important for stability of ecosystem function in a changing environment. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 371, 20150272. 10.1098/RSTB.2015.0272 [DOI] [PMC free article] [PubMed] [Google Scholar]
  97. Yang, Q. , Zhu, W. , Tang, X. , Wu, Y. , Liu, G. , Zhao, D. , Liu, Q. , Zhang, Y. , & Zhang, T. (2024). Improving rice grain shape through upstream ORF editing‐mediated translation regulation. Plant Physiology, 197, kiae557. 10.1093/PLPHYS/KIAE557 [DOI] [PMC free article] [PubMed] [Google Scholar]
  98. Yin, L. , Zhang, H. , Tang, Z. , Xu, J. , Yin, D. , Zhang, Z. , Yuan, X. , Zhu, M. , Zhao, S. , Li, X. , & Liu, X. (2021). rMVP: A memory‐efficient, visualization‐enhanced, and parallel‐accelerated tool for genome‐wide association study. Genomics, Proteomics and Bioinformatics, 19, 619–628. 10.1016/j.gpb.2020.10.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  99. Zhou, X. , & Huang, X. (2019). Genome‐wide association studies in rice: How to solve the low power problems? Molecular Plant, 12, 10–12. 10.1016/J.MOLP.2018.11.010 [DOI] [PubMed] [Google Scholar]
  100. Zou, S. , & Kumar, U. (2018). Cannabinoid receptors and the endocannabinoid system: Signaling and function in the central nervous system. International Journal of Molecular Sciences, 19, 833. 10.3390/IJMS19030833 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S1: Description of cannabinoid traits in the 174 drug‐type Cannabis accessions. Data were provided by Lapierre et al. (2023). THCV, CBDV and CBN are qualitative variables, with 0 indicating absence and 1 indicating presence.

Table S2: Genome‐ and chromosome‐wide nucleotide variation analysis across the GWAS panel, clustering assignments and chemotypes. The 154 Cannabis accessions were distributed as K1 = 35, K2 = 52 and K3 = 87, Type I = 152, Type II = 17 and Type III = 5.

Table S3: Identification of candidate genes through haplotype block (HB) analysis and functional impact analysis of SNPs. Orthology analysis of Cannabis genes was performed against the Arabidopsis thaliana transcriptome TAIR 11 (Cheng et al., 2017). Only significant SNPs and SNP residing in HB of interest located in a gene sequence were considered for functional impact analysis (detailed in Supplemental Table 3).

Table S4: Functional variant analysis of significant SNPs and SNP residing in HB of interest located in a gene sequence. The SNPs having a putative impact on gene function were bolded.

Figure S1: Heatmap of pairwise kinship matrix values between the 174 drug‐type accessions. The color histogram showsthe distribution of co‐ancestry coefficients. The redder the color, the more similar the individuals are, whereas the more yellow the color, the more dissimilar the individuals are.

Figure S2: Discriminant analysis of principal components (DAPC) of the 174 drug‐type accessions using the whole 282K high‐quality SNPs. a. Cumulated variance explained by the eigenvalues of the PCA. b. Value of BIC versus for increasing values of k. Cluster selection was based on the BIC value. c. Optimization α‐score graph. The optimal α‐score suggested that twenty PCs should be retained for the discriminant analysis. d. DAPC cross‐validation test for the optimal number of PCs retained.

Figure S3: Population structure analysis of the 174 drug‐type accessions using 282K high‐quality SNPs. Dark blue, light blue and green correspond to cluster K1, K2 and K3, respectively. a) Admixture plot for K = 3 using fastStructure. Each vertical line represent an accession. b) Discriminant analysis of principal components (DAPC) scatter plot. c) Unrooted phylogenetic tree estimated using the Unweighted Pair Group Method with Arithmetic Mean (UPGMA). Colors are based on fastStructure clustering. d) Concordance of cluster assignment for fastStructure and DAPC inference.

Figure S4: Quantil‐quantil (QQ) plots for eleven cannabinoid traits in drug‐type Cannabis.

Figure S5: Distribution of allelic classes of significant markers associated with cannabinoid traits in 174 drug‐type Cannabis accessions.

Figure S6: Distribution of allelic classes of 5 SNPs with putative impact on gene function.

Figure S7: Distribution of allelic classes of significant SNPs compared to chemotype categories. Chemotype are based on the THC/CBDratio. Major alleles are highlighted in blue, minor alleles in orange, and heterozygous alleles in grey.

TPG2-18-e70031-s001.xlsx (41.4KB, xlsx)

Data Availability Statement

The sequencing files generated used for the analyzes of this study are on NCBI BioProject: PRJNA1047573 and will be accessible after acceptance of the manuscript (reviewer link: https://dataview.ncbi.nlm.nih.gov/object/PRJNA1047573?reviewer=kr7i954sp11ttcn4qm36trajc4).

Articles from The Plant Genome are provided here courtesy of Wiley

RESOURCES