Plasticity, allelic diversity, and genetic architecture of industrial hemp (Cannabis sativa L.)

Abstract

The first time in United States history that hemp was legally distinguished from high-THC Cannabis (marijuana) was in 2014 when the Farm Bill was passed. Although the two crops had been distinguished by their usage for thousands of years, their monospecific nature led to both psychoactive and non-psychoactive forms being legislated in tandem from the time that Americans began regulating Cannabis cultivation and usage. A simple statement in the 2014 Farm Bill distinguished hemp as Cannabis sativa L. with a tetrahydrocannabinol (THC) content of 0.3% or less. A second sentence enabled research into the crop and production within pilot programs in states where it is legal. This minor change in legality, followed by subsequent relaxation of laws surrounding hemp in the 2018 Farm Bill, has allowed a burgeoning hemp industry to form in the United States and enabled the return of a relict crop. Due to the long period of prohibition, hemp did not undergo the same type of crops research as other staple American crops. Consequently, little is known about the genetic mechanisms that control many of the key traits in hemp production. Understanding basic information about how traits are affected by environmental factors is highly important when regulation of the crop is based on a stringent and arbitrarily set threshold for chemical content. In 2016, we performed field trials of a diverse set of industrial hemp cultivars in multiple growing environments and assessed a wide range of traits. Expression of some traits, like days to maturity and THC content, were strongly influenced by genotype. Other traits, such as grain yield and plant height, exhibited large proportions of variance due to environmental factors and genotype-by-environment interactions. There were also varying ranges of plasticity exhibited between cultivars, underscoring the importance of selecting the right cultivar for target production environments. This highlights the importance of thoroughly characterizing genotype-by-environment interactions when breeding locally adapted hemp cultivars. Understanding genetic control of important traits and their range of plasticity enables the development of locally adapted cultivars for a wide range of end uses. Another aspect of Cannabis that is understudied is the genetic basis for differentiating hemp and high-THC Cannabis. Since the legal distinction is based on a strict threshold placed on a quantitative trait and not any known geographic or biological reproductive barriers, it is unclear whether or not there is genetic evidence to support the distinction or if the two groups are simply divergent phenotypes. A joint-site frequency and FST analysis show that individuals of the two groups mainly share common polymorphisms, with a small number of loci where differentiation occurs. These loci serve as the basis for distinguishing the two groups, but more study is needed to determine if alleles in these regions were driven to fixation via genetic drift and selection on unrelated traits, or if there is an evolutionary basis for the observed differences. When heterozygosity was assessed in these samples, the hemp group had higher overall heterozygosity levels, but the high-THC Cannabis group had more outliers which lead to a wider distribution with more extreme minimum and maximum values. Although it is clear that there are genetic differences distinguishing the two groups, extensive human vectoring and admixture between the groups, both historically and currently, makes it difficult to differentiate causes for the differences. A lack of centralized germplasm makes large-scale genomic studies of the species difficult, but, as more samples are surveyed over time, a more detailed picture of the genomic variation will emerge. These types of studies will be able to provide a more nuanced picture of the evolutionary history and current state of allelic variation within the species. In addition to plasticity and allelic diversity, genetic architecture of traits has also largely been ignored until recently. The first QTL study in Cannabis was performed in 2015 and was limited by legal restraints. Since understanding how economically relevant traits function is important to breeding improved hemp cultivars, we developed a genetic mapping population that captured variation for a wide range of traits. Utilizing whole-genome sequencing and phenotype data from a replicated field trial, we were able to detect 121 QTL associated with 38 agronomic and biochemical traits. Some traits, like days to maturity, had single loci of large effect accounting for the majority of trait variance, while other traits, like α-Pinene production, exhibited more complex polygenic architecture with epistatic interactions. Colocalization of QTL and significant trait correlations showed that there were positive relationships within both agronomic and biochemical trait groups. Although this study was limited by assessment of the population in a single environment, detecting these putative QTL serves as a substantial step forward in characterizing many relevant production traits.