Repetitive sequences occupy the majority of any typical eukaryotic genome, yet are poorly understood in many respects. The effect of this repetitive content has been under debate for decades [1]. Numerous scientists regard it as parasitic or ‘selfish DNA’ [2]. Others think repetitive elements might play important roles in the host’s genome by altering a gene’s function [3], or by acting as raw material for new genes [4]. These ideas are not mutually exclusive, with repetitive sequences likely having both positive and negative effects in most genomes. Repetitive DNA elements can be mainly classified into two major groups based on their organization in the genome [5]. One group includes sequences showing a tandem repeat organization, where copies are arranged adjacently to each other, commonly (though not always) in or near centromeric and telomeric regions [5, 6]. DNA elements which form tandem arrays such as satellite DNA, simple repeats and ribosomal DNA (rDNA) occur primarily in tandem repeat blocks [5, 7]. A second group of repetitive DNA sequences consist of elements which are dispersed across the whole genome [8]. These include mobile elements like DNA transposable elements (TEs) and retrotransposons such as long terminal repeat elements (LTRs), short interspersed nuclear elements (SINEs), and other dispersed repeats [9, 10].

Eukaryotic species show huge variation in genome size, ranging from only 2.3 Mb in the microsporidian Encephalitozoon intestinalis [11] to over 152 Gb in the plant Paris japonica [12], five orders of magnitude of variation largely due to changes in repetitive content of the genome, although changes in gene content and ploidy also play a role [13]. Repetitive elements occupy a substantial fraction in most plant genomes, ranging from 20% in Arabidopsis thaliana to more than 80% in Helianthus annuus (sunflower) [14, 15]. Copy numbers of mobile elements range from thousands to millions per diploid genome [16]. Indeed, just within flowering plants, genome sizes differ roughly by 2500-fold largely due to variation in the copy numbers of TEs and other highly repetitive sequences [17, 18].

The most common repeat sequences fall into several major categories. Due to their copy-paste transposition mechanism, active retrotransposons have the potential to increase their copy number affecting the genome size. LTR/Gypsy and LTR/Copia are two super-families of retrotransposons present in high copy numbers, resulting in major fractions in flowering plant’s genomes [10]. Usually a smaller proportion of the genome, but highly important, ribosomal DNA (rDNA) encoding sequences in higher plants are arranged in long tandem arrays [7]. Plants typically have 500 to 40000 rDNA copies per diploid cell [19]. Simple sequence repeats, which exhibit high mutation rates, are also abundant within both animal and plant genomes [20]. Together, these classes of sequences make up the majority of the high-copy sequences in most well-characterized eukaryotic genomes.

Some classes of repetitive elements, such as LTR elements, provide an opportunity for deciphering the evolutionary demography of a family of retrotransposons within host genomes. Newly produced retrotransposons are 100% identical to the parental molecule but with no mechanism to maintain their homogeneity after insertion, they are expected to diverge neutrally [18]. Mutations gradually disfigure the elements to different lengths leading to an incomplete structure that might also inactivate them [21]. The magnitude of pairwise divergence between two LTRs can be used to infer their relative ages [22]. Studies on LTR families in rice [23], maize [22] and peas [24] calculated insertion ages using pairwise divergence among the elements and found that the average level of divergence for a large fraction of LTR elements is on the order of 1% or less [18]. These genomes thus are largely composed of recently duplicated repeat sequences.

Cannabis sativa (marijuana, hemp), a member of the family Cannabaceae is a widely cultivated plant with numerous genomic questions unanswered. The family Cannabaceae has 11 genera with approximately 100 species widely distributed throughout Asia and Europe [25]. Cannabis sativa is one of the earliest domesticated and cultivated plant species for fibre, oil, and for its medicinal and psychoactive properties [26]. Cannabis sativa has a diverse set of metabolic compounds as cannabinoids and terpenoids [27] and the numerous varieties vary morphologically in the production of these compounds [28]. Cannabis sativa Purple Kush (PK) is a commonly used recreational marijuana variety, while the cultivars ‘Finola’ and ‘USO31’ are hemp varieties used industrially [29]. These three varieties were sequenced in 2011 [30] and their sequences are publicly available at NCBI’s short read archive. The estimated size of the haploid C. sativa genome is 830 Mb and the draft genome assembly contains approximately 80% of the estimated genome [30]. In order to understand more about the genome of this important species, we studied its genomic diversity by exploring the TE dynamics.

Humulus lupulus (hops), the closest relative of C. sativa, has a genome three times as big (2570 Mb) [31]. Morus notabilis (mulberry), a closely related species to both C. sativa and H. lupulus diverged approximately 63.5 MYA [32] and was the closest related species of C. sativa sequenced until recently. The estimated genome size of M. notabilis is 357 Mb, which is less than half of C. sativa’s genome [32]. Thus, these genera span a relatively large range of genome size variation. Using the genomes of these three related genera allows us to make comparisons between them to explain the variation in genome size within and among species of Cannabaceae family.