Source: UNIV OF WISCONSIN submitted to
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
Funding Source
Reporting Frequency
Accession No.
Grant No.
Project No.
Proposal No.
Multistate No.
Program Code
Project Start Date
Sep 1, 2021
Project End Date
Aug 31, 2024
Grant Year
Project Director
Ellison, S.
Recipient Organization
21 N PARK ST STE 6401
MADISON,WI 53715-1218
Performing Department
Non Technical Summary
The promise of hemp, with thousands of possible uses, potential suitability to numerous environments, and emerging, fast-growing new product markets, could be a welcome alternative crop for many producers. However, despite significant developments in the U.S. hemp industry over the past several years, there remain substantial gaps in the knowledge needed to fully support this fledgling industry. A recent survey of over 1,100 hemp stakeholders found that over 75% think additional research in breeding and genetics to produce stable and uniform cultivars and regional adaptability is very or extremely important. This is directly related to seed characteristics and considerations such as germination rate, dormancy, shelf life, state certification, cost, and quality. Genetic improvement of cultivars will have a direct effect on yield and production cost and therefore the return on investment for stakeholders. In order for public and private breeders and scientists to develop regionally adapted hemp cultivars it will be necessary for them to have access to a broad range of germplasm. Initial characterization of hemp germplasm shows a high level of genetic diversity in naturalized hemp populations compared to cultivated Cannabis, likely due to selection pressures and genetic drift in different regional populations. The purpose of this research task is to build and characterize a genetically diverse compliant collection of feral hemp populations that are adapted to environmental conditions in the regions where they have been growing since escaping from agronomic cultivation. Through the engagement and use of citizen scientists and regional collaborators we will be able to cover more ground than any small group of individuals. This collection will serve as a public resource for the characterization of genes and mining of alleles for beneficial traits in breeding new cultivars.
Animal Health Component
Research Effort Categories

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
Knowledge Area
202 - Plant Genetic Resources;

Subject Of Investigation
1730 - Hemp;

Field Of Science
1081 - Breeding; 1080 - Genetics;
Goals / Objectives
Overall Goal: Our integrated research and extension goal is to collect a geographically distributed range of genetically diverse and THC-complaint hemp germplasm that will kick start research and breeding efforts, while collaborating with and educating citizen scientists and stakeholders about industrial hemp and its potential downstream uses.Specific ObjectivesObjective 1: Collect seed and passport data from feral hemp populations representing the diversity of ecosystems in the United StatesObjective 2: Characterize cannabinoid traits for regulatory compliance and seed bankingObjective 3: Distribute a core collection of compliant discovery populations to collaborators for trait identification and breeding efforts
Project Methods
Obj 1 Experimental approach:Population identification-We will enlist citizen scientists to identify, verify, and collect site specific data for geographically distributed feral hemp populations. Specifically, we will use the iNaturalist application, which is free, easy to use, and one of the most popular nature applications with over one million registered users.In addition to using the citizen scientist approach, we have identified key regional partners who will assist with regional documentation of feral populations and collections. These partners will work with their local governments, state agencies, farmer groups, and nature groups to identify additional populations.Population collection-The country will be divided into geographic regions of interest based on the National Health and Environmental Effects Research Laboratory U.S. Environmental Protection Agency Level III Ecoregions. These ecoregions are areas where ecosystems and the type, quality, and quantity of environmental resources are generally similar. Regional collaborators have been selected based on geographic distribution and will assist with collection efforts in the late summer of 2022 and 2023. We will target ~50 ecoregions with 5 populations per region for a total of 250 geographically dispersed populations. All collaborators are familiar with their state licensing policies and must have a license for each year they are involved with collections. We will provide a collection protocol and list of supplies for each collaborator. The protocol will consist of harvesting seed from up to 50 female plants per population.Analysis:Passport data collected via iNaturalist and from regional collaborators will use standard Multi-Crop Passport Descriptors (MCPD) which has been developed as a uniform, global format. Analysis of materials is described further in Objective 2 and 3.Obj 2 Experimental Approach:Genetic Testing-We propose to pre-characterize every seed source (maternal plant) individually and a large random sample of seed from hemp-type maternal plants for each population. Testing maternal parents excludes drug-type genetics from the maternal line (dam). Seedling tests can estimate the frequency of the drug-type allele in the population using the paternal line (sire). Taken together, the two levels of testing can identify the best candidate populations for seed banking as well as populations of special concern.A single leaf taken from each female plant for which seed is harvested, stored separately in a coin envelope and individually indexed to the seed lots will be genotyped. Genomic DNA from the leaf sample of each female plant will be isolated using a CTAB protocol adapted from (Doyle and Doyle 1987). A portion of the gDNA will be used to determine the CBDAS genotype of the maternal field plants using the low cost manual CAPS assay described and validated by Wenger et al. (2020). One individual, determined by the cannabinoid profiling to be CBD-type and verified by the CAPS assay to be homozygous for functional CBDAS, will be selected as seed source for further genetic characterization. Approximately 5 ug of the isolated gDNA from each of the field collected plants from each site will be shipped to PI Ellison at UW-Madison for GBS analyses (see Objective 3).From the selected CBD-type individual for each collection site, we aim to germinate 100 seeds and maintain seedlings for two weeks of vegetative growth. True leaves of each seedling will be harvested, dried in the lab at ambient conditions, and submitted to the University of Minnesota Genomics Center (UMGC) where gDNA will be isolated after bead-shake pulverization using a CTAB protocol adapted for high-throughput (HTP). From gDNA, the CBDAS genotype of each seedling will be scored at UMGC using Illumina-based targeted amplicon sequencing.Cannabinoid Testing-Cannabinoid screening of a subset of seed heads for each population will serve several purposes. First, cannabinoid profiles will maintain quality control of the genetic assay. Second, the CBD:THC ratio test is insufficient to measure overall cannabinoid content (TCC). Third, we will have comparative data on variation in cannabinoid levels across the geographic distribution of source populations. Lastly, field measures of THC are required for state and federal compliance while aiding local collaborators and officials in decision-making should non-compliant populations be identified as a public safety concern.From each location from which feral seed is collected, seeded inflorescences (colas) from a minimum of ten individual female plants (10 plants per site where N = 50, otherwise 20% of N) will be reserved as indicated in Objective 1. Seeded inflorescences will be dried by collectors and shipped to UMN for determination of cannabinoid content of six cannabinoid compounds (CBC, CBD, CBG, d8-THC, d9-THC, THCV) using GC-MS. All cannabinoid phenotyping is performed in the blind with respect to provenance and sample codes to minimize the potential for investigator bias. Seed from each individual plant inflorescence will be separated from maternal tissue and held individually to allow subsequent genetic analyses of seed informed by maternal cannabinoid phenotype (THC:CBD ratio and TCC).Data Analysis and Application-For plants collected in the field by regional collaborators, CBDAS locus genotypes will be called by direct inspection from CAPS assay gel experiments (Fig. 1B, Wenger et al. 2020) and chemotypes will be determined using % dry weight inflorescence fractions as x = log(%THC/%CBD) where values of x >= 1 are THC-type, x =< -1 are CBD-type, and 1 > x > -1 are intermediate-type. CBDAS genotypes of seedlings will be called by UMGC from Illumina-based targeted amplicon sequencing using a marker assay (primer set) validated against the Weiblen laboratory's THC-type x CBD-type F2 mapping population (Weiblen et al. 2015). CBDAS genotype and allele frequencies of seedlings will be calculated using JMP Pro 14.2.0 (SAS, Cary, NC, USA) and tests of departure from Hardy-Weinberg equilibrium will be performed using a likelihood ratio test implemented in ExactoHW 1.1 software (Engels 2009).Obj 3 Experimental approach:We will acquire gDNA from Objective 2 from three of the original collected female plants per population that proved compliant from both cannabinoid screening and CBDAS genotyping. The gDNA will be genotyped-by-Sequencing on an Illumina NovaSeq 6000 at the University of Wisconsin Biotechnology Center DNA Sequencing Facility. Single nucleotide polymorphisms will be called using the GBS TASSEL pipeline using the Cannabis sativa reference genome CBDRx assembly (project PRJEB29284) (Glaubitz et al. 2014, Grassa et al. 2021). SNPs will be filtered for quality, depth, and minor allele frequency.Analysis:To better understand genetic relatedness, a series of analyses will be conducted including population structure (Alexander and Lange 2011), phylogenetic relatedness (neighbor-joining tree in PHYLIP), Principal Component Analysis (SNPRelate), Nei's gene diversity, Shannon's information index, and polymorphism information content (PICcalc). The consensus number of subpopulations will be used for AMOVA and Nei's genetic distance, fixation index (Fst). In addition, genetic indices such as number of loci with private allele, number of different alleles (Na), number of effective alleles (Ne), Shannon's information index (I), observed heterozygosity (Ho) and expected heterozygosity (He) will be calculated with the R package Poppr (Kamvar Z. N. 2014). Core collections will be made using Core Hunter 3 ( project will be evaluated based on the number of collaborators (20+), citizen scientists (+50), ecoregions coverage (+50), hemp samples obtained (+100), hemp samples distributed to researchers (+50), and hemp samples distributed to the hemp germplasm repository (+100).

Progress 09/01/22 to 08/31/23

Target Audience:We have reached several dozen citizen scientists over iNaturalist to help identify and collect feral hemp populations. Citizen scientists were educated through emails and virtual meetings about the importance of making these collections to preserve germplasm and fill the new hemp seed bank to initiate research and breeding programs. We also presented our work to fellow scientists through conferences and symposia to describe the current state of collections and enlist new collectors for the upcoming year. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The postdoctoral researcher working on this project was able to receive collaborative research opportunities by working with another lab to gain exposure to the chemotyping assay. They then brought that back to our lab and were able to train others and gained skills in mentoring. They were also able to present preliminary findings a research symposium and field day, gaining skills in scientific communication for both scientific and general audiences. They will also be attending a week long bootcamp on high throughput computing in August 2023. Undergraduate researchers in the lab were trained in molecular biology techniques and data analysis. Citizen scientists received one-on-one training from the funded postdoc to collect feral hemp accessions. Meetings usually occurred over Zoom and the collection protocol was discussed as well as emphasizing the importance of the work and their contribution to germplasm preservation. How have the results been disseminated to communities of interest?Our preliminary results were presented a several symposia and conferences including the 2022 UW Madison Hemp Research Symposium and the Cannabis Research Conference - Denver, CO. The project will also be showcased at the UW Madison Hemp Research Field Day - Arlington, WI on August 17th, 2023 and UW Madison Plant and Agroecosystem Sciences and Soils Arlington Field Day - Arlington, WI on August 30th, 2023. The project was also discussed during guest lectures at various cannabis courses for undergraduates at UW Madison, Gateway Technical College, and Delaware Valley University. Student learned about the importance of germplasm preservation and some even enlisted friends and family members to help make seed collections. What do you plan to do during the next reporting period to accomplish the goals?During the late summer and fall of 2023 we will again enlist our collection collaborators to obtain feral hemp samples across the country. These samples will be sent to Dr. Weiblen's group where his team will test flower for cannabinoid content using HPLC and extract DNA from leaves. The DNA will be used to determine chemotype and the remaining DNA will be sent to Dr. Ellison's group for Genotype by Sequencing (GBS). Seeds will be considered "hemp" or "drug-type cannabis" and submitted to Dr. Zac Stansell for inclusion into the USDA hemp germplasm collection. After all feral hemp has been genotyped, samples will be analyzed using population genomics pipelines to determine genetic relatedness. A core collection of genetically unique samples will be redistributed to collectors to initiate regional breeding efforts. A planned submission to Advances in Conservation and Utilization of Plant Genetic Resources Special Edition - Frontier in Plant Science - is also anticipated to disseminate our research project protocol and findings to interested communities.

What was accomplished under these goals? The purpose of this research task is to build and characterize a genetically diverse collection of feral hemp populations that are adapted to environmental conditions in the regions where they have been growing since escaping from agronomic cultivation almost a century ago. We have already been able to successfully collect over a thousand feral hemp samples and find there is both significant genetic and phenotypic diversity across this collection. Additionally, we find most of these materials are THC compliant and they will be able to be immediately used by the hemp research community. These samples will serve as a resource for the characterization of genes and mining of alleles for beneficial traits in breeding new hemp cultivars. The genetic improvement of hemp cultivars will provide farmers with another option to diversify their current cropping system to improve their economic and environmental options and outcomes. Objective 1: Collect seed and passport data from feral hemp populations representing the diversity of ecosystems in the United States Collection and Sequencing One thousand and fifty (n = 1,050) feral accessions were collected from eight midwestern states in the late summer and early fall of 2022. Objective 2: Characterize cannabinoid traits for regulatory compliance and seed banking Characterization of Cannabinoid Traits Field-collected feral flower samples were chemotypedfor cannabinoids using PACE genotyping protocol (Toth et al., 2020). The majority of the sampled feral population across the Midwest were compliant with CBD-type alleles while one population show a fair amount of all three groups (CBD-type, Intermediate, and THC-type). Five to 10 flower samples from each collection were submitted to U. Miss for HPLC in May 2023. Objective 3: Distribute a core collection of compliant discovery populations to collaborators for trait identification and breeding efforts Population Structure and Genetic Diversity Statistics A subset (n = 453) of samples were used for DNA analysis. Genotyping by Sequencing (GBS) was employed for single nucleotide polymorphism (SNPs) marker development of a total of 98,298K SNPs. With a 0.05 minor allele frequency and a minimum call rate of 0.1 (i.e., no more than 10% of the genotyped feral accessions in the germplasm panel should have missing data), ~35K SNPs were used for further analysis. Data were analyzed using software ADMIXTURE (Alexander et al., 2009). Discriminant analysis of principal components and principal component analysis were performed using adegenet package in R (Jombart et al., 2010; Jombart and Ahmed, 2011). Analysis of molecular variance was also calculated with poppr package in R (Kamvar et al., 2014). Nei's genetic distances and Fst statistics were calculated using hierfstat package (Goudet 2005; Goudet and Jombart, 2015). These initial results shows significant genetic diversity was present in our first-year collection of feral hemp samples and three major clusters were identified. Principal component analysis indicated that PC1 and PC2 explained 15.7% and 9.21% of the genetic variation among the accessions. Admixture analysis revealed three populations, one unique Indiana population, and two additional subpopulations. One of the subpopulations consisted mostly of WI and IL feral hemp while the other had mostly NE and IA feral hemp. Discriminant analysis of principal component results corroborated Admixture results with a similar clustering patterns. Fixation indices values between populations (> 0.15) indicated that the populations were significantly different from one another. AMOVA results also indicated variation within samples, between populations, and between samples within a population.


    Progress 09/01/21 to 08/31/22

    Target Audience:We have created an iNaturalist Project to recruit citizen scientists to upload feral hemp locations. The global iNaturalist community has over 2.5 million registered users as of December 2021. We have also reached out to feral hemp collection collaborators via emails and virtual presentations to provide training. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?A postdoctoral researcher was hired to assist with collections, genotyping samples, and redistribution of materials. They will be trained in bioinformatics, genomics, database management, and scientific communication. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals?There is very little hemp germplasm available to initiate a domestic hemp breeding program. In the next year, we intend to collect a wide geographic distribution of feral hemp. This material will be characterized for compliancy (THC < 0.3%) and deposited into the USDA Hemp seed bank for preservation, increase, and distribution. The feral hemp collections will also be genotyped to determine the most unique samples to reduce the need to spend resources on similar samples. This material will be redistributed to breeders to integrate useful new genetic diversity to increase yield. Increased yield will make hemp a more viable crop for farmers to grow profitably.

    What was accomplished under these goals? As of today, there is no publicly available hemp germplasm in the United States. It is likely the efforts described in this project will lead to the identification of thousands and collection of hundreds of novel feral hemp populations. This material will be critical in ensuring the success of hemp production in various ecoregions across the United States. Objective 1: Collect seed and passport data from feral hemp populations representing the diversity of ecosystems in the United States A sampling protocol was developed and field tested. The protocol has been presented to feral hemp collectors in virtual training sessions and included in the USDA hemp descriptor and phenotyping handbook. Objective 2: Characterize cannabinoid traits for regulatory compliance and seed banking Nothing to report for this period. Objective 3: Distribute a core collection of compliant discovery populations to collaborators for trait identification and breeding efforts Nothing to report for this period.