Introduction

Wheat (Triticum aestivum), barley (Hordeum vulgare), rye (Secale cereale), and the wild relatives of these crops (e.g., members of the genus Aegilops) are closely related and belong to the Triticeae tribe, which evolved some 12 million years ago within the Pooideae subfamily of the Poaceae (Gaut 2002). The haploid genome of diploid Triticeae is ca. 5 Gbp, as estimated based on the genome of diploid cultivated barley (IBSC 2012); over 80% of this genome comprise repetitive elements. All three genomes (A, B, and D) of common hexaploid wheat are similar to the genome of barley in terms of genome size, gene content, and repetitive elements (IWGSC 2014; Wicker et al. 2018). Barley and wheat genes are functionally closely related. Therefore, information about cloned barley genes associated with a particular trait can be used to help identify the genes responsible for a similar trait in wheat. This was clearly demonstrated by a study involving genome editing of three wheat orthologs (Abe et al. 2019) of the barley seed dormancy gene Qsd1 (Sato et al. 2016), leading to high levels of seed dormancy in the genome-edited wheat plants.

Several years after its establishment, the International Sequencing Consortium released draft genome sequences for barley (IBSC 2012), Aegilops tauschii (Luo et al. 2013), and common wheat (IWGSC 2014). These draft genomes were based on the physical maps of BAC (bacterial artificial chromosome) clones covering entire chromosomes and on whole-genome shotgun sequencing assemblies of short reads with gene model annotations based on transcript evidence. Overlapping BAC clones were then arranged into minimum tiling paths, which were subjected to shotgun sequencing. Based on these efforts, high-quality chromosome-level assemblies were released for barley (Mascher et al. 2017), A. tauschii (Luo et al. 2017), and common wheat (IWGSC 2018). Multiple haplotypes of both barley (Jayakodi et al. 2020) and wheat (Walkowiak et al.2020) were recently de novo sequenced, revealing considerable genomic variations within each crop species.

The wheat gene pool encompasses a large number of species in the genera Triticum and Aegilops, including the donor species of hexaploid-cultivated wheat. The availability of chromosome-scale genome assemblies for wheat and barley has facilitated genetic and genomic analysis of cultivated and wild wheat species by providing the basic structures of pseudomolecules and genes. However, the sequences of wild wheat genomes other than the donor genomes of cultivated wheat remain uncharacterized. Genetic and evolutionary analyses of wild wheat species are particularly challenging due to dissimilarities in the structures and sequences of the assembled genomes.

The wheat and barley research communities have generated numerous cDNA sequences that can be used to characterize the expressed portion of the genome and provide a source of genetic markers (Close et al. 2009; Manickavelu et al. 2012). Following the development of next generation sequencers, it became possible to perform massive RNA sequencing (RNA-Seq) of cDNA libraries for expression analysis and the detection of polymorphisms among samples. Many RNA-Seq studies have been performed for expression analysis. Xiang et al. (2019) used this method to examine common and durum wheats and their three putative diploid ancestors and identified gene expression programs and the contributions of the A, B, and D subgenomes to grain development in polyploid wheats. Nevertheless, identifying polymorphisms between polyploid wheat haplotypes remains challenging because short transcript sequences from multiple (sub)genomes were assembled within gene models. Fortunately, RNA-Seq analysis can be used to efficiently detect polymorphisms even among closely related diploid germplasms, such as the breeding germplasms used for barley (Tanaka et al. 2019).

Here, we review our use of RNA-Seq analysis to detect polymorphisms in diploid wheat genomes and generate DNA markers. Analysis of the barley genome often provides useful information about the structures of diploid Triticeae genomes due to the well-annotated gene models and clear descriptions of genome structures available for barley.

DNA marker generation in A. tauschii and its application to hexaploid wheat analysis

A. tauschii is an important D-genome donor species of common wheat. The D-genome of common wheat is less polymorphic than its A- and B-genomes (Rosyara et al. 2019). Tetraploid wheat and A. tauschii can be crossed artificially to produce synthetic hexaploid wheat (Kihara and Lilienfeld 1949; Matsuoka and Nasuda 2004) (Fig. 1a). These synthetic lines can be used as intermediates to exploit the natural variation in A. tauschii for hexaploid wheat improvement (Trethowan and Mujeeb-Kazi 2008). However, compared to cultivated wheat species, the analysis of natural variation in A. tauschii has been limited by the lack of efficient genetic markers for this wild wheat species. The main PCR-based marker system that has been used until recently is simple sequence repeats (SSRs; Somers et al. 2004), but the marker resolution for this system is low, and the genomic positions of SSRs are unknown.

Fig. 1
figure 1

Relationships among Triticeae species. a Images of spikes of hexaploid synthetic wheat (AABBDD) and parental lines Triticum turgidum (AABB) and Aegilops tauschii (DD). b Molecular relationships among diploid Triticeae species (Tanaka et al. 2020)

Full size image

By contrast, the genomic positions of transcript-based markers can be estimated. Iehisa et al. (2012) and Iehisa et al. (2014) used the early next generation sequencing (NGS) platform known as the 454 system (GS FLX Titanium, Roche Diagnostics) and RNA-Seq to detect polymorphisms among representative A. tauschii accessions (Table 1). Based on an analysis of the population structure of A. tauschii accessions and amplified fragment length polymorphism analysis of 122 accessions, Mizuno et al. (2010) identified two major phylogenetic lineages: 1 (L1) and 2 (L2). Iehisa et al. (2012) used RNA-Seq to analyze two A. tauschii accessions from each lineage. Following de novo assembly of the reads, 9435 single-nucleotide polymorphisms (SNPs) and 739 insertion/deletion polymorphisms (indels) were identified.

Table 1 Summary of RNA-Seq data for Triticum and Aegilops species
Full size table

Iehisa et al. (2014) developed markers by deep sequencing the spike transcriptomes from the same two A. tauschii accessions and mapping the reads to the F2 population of these accessions. The authors then transferred the marker information to hexaploid wheat accessions, including synthetic hexaploid wheat lines. The synthetic wheat lines were obtained through crosses between tetraploid wheat cv. Langdon and the A. tauschii accessions, followed by chromosome doubling of the interspecific hybrids (Takumi et al. 2009; Kajimura et al. 2011). In total, 5642 out of 5808 contigs with high confidence (HC) SNPs were assigned to the A. tauschii draft genome sequence (Luo et al. 2013). Iehisa et al. (2017) detected splicing variants in the transcripts of two A. tauschii accessions and applied their splicing patterns to synthetic hexaploid wheat. They analyzed alternative splicing of 23,778 loci and identified multiple splicing variants in 4712 genes. For at least some genes, alternative splicing patterns were clearly distinct between the two A. tauschii accessions and were transmitted from the parental A. tauschii accessions to the synthetic hexaploid wheat lines.

Nishijima et al. (2016) generated additional markers using the more recent MiSeq (Illumina) NGS platform to generate 300-bp paired-end reads, which overlapped with the target fragment size of 550 bp. The number of reads ranged from 669,383 (247 Mbp) to 893,917 (386 Mbp) using the 454 platform with 500-bp single reads (Iehisa 2012; Iehisa 2014) but from 4.8 to 5.8 million using MiSeq 300-bp paired-end sequencing reads. After de novo assembly of the merged paired-end reads of ten representative accessions of A. tauschii, 33,680–65,827 transcripts were reconstructed for each accession, with N50 values of 1369–1519 bp. This analysis formed the basis for generating RNA-Seq transcripts, providing a set of unigenes for each accession, including 29,386–55,268 representative isoforms. The deduced unigenes were anchored to the chromosomes of A. tauschii and barley.

The SNPs and indels in the anchored unigenes, which covered entire chromosomes, were sufficient for linkage map construction, even when using combinations between the most closely related accessions. Interestingly, the resolution of SNP and indel distribution was higher in the barley genome than in the A. tauschii genome. Since barley chromosomes are regarded as virtual chromosomes of Triticeae species, this strategy allows genetic markers to be generated that are arranged on the chromosomes in an order based on conserved synteny. This technique is efficient for genome-wide RNA-Seq-based DNA marker generation for A. tauschii.

Applications of DNA markers to the genetic analysis of A. tauschii

The ability to generate markers by RNA-Seq analysis allowed us to use these markers for genetic analysis. Bulked segregant analysis (BSA) based on RNA-Seq is a powerful method for narrowing the search for candidate causal genes of target phenotypes in polyploid wheat and its wild relatives (Trick et al. 2012; Ramirez-Gonzalez et al. 2015; Edae and Rouse 2019; Li et al. 2019). Nishijima et al. (2018) performed BSA of a bi-parental mapping population of two synthetic hexaploid wheat lines that shared identical A and B genomes but contained D genomes of distinct origins with different Net2 alleles (Fig. 2). An F2 mapping population was generated via a cross between these hexaploid lines. Heterozygotes in subsequent generations were selected and used to generate two F5 lines: one with ten plants homozygous for Net2 and the other with ten plants homozygous for net2. Using the platform described by Nishijima et al. (2016), pooled RNA samples were analyzed by RNA-Seq-based BSA, and D-genome-specific polymorphisms around the Net2 gene were identified. The resulting SNPs were classified into homoeologous polymorphisms of the A, B, and D genomes and D-genome allelic variations based on the RNA-Seq results from a parental tetraploid and two A. tauschii accessions. The difference in allele frequency at the D-genome-specific SNP sites between the contrasting bulks (the ∆SNP-index) was higher for the target chromosome than for the other chromosomes. Several SNPs with the highest ∆SNP-indices were converted into PCR-based markers and assigned to the Net2 chromosomal region, which will subsequently be narrowed down to a single gene.

Fig. 2
figure 2

modified from Nishijima et al. (2018)

An example of RNA-Seq based bulk segregant analysis for locus mapping in wheat. RNA-seq reads generated by Illumina MiSeq were aligned to A. tauschii cDNAs, and SNP calling was conducted for SNP-index calculation to identify SNP markers linked to the locus. This figure was

Full size image

RNA-Seq-based markers are also being used to detect segments introgressed from A. tauschii to a synthetic hexaploid wheat line showing early heading as a novel source of variation. Takumi et al. (2020) backcrossed two early heading lines of a synthetic hexaploid wheat line derived from a cross between durum wheat and A. tauschii with four Japanese elite cultivars to develop early heading lines of bread wheat. For this technique, it is important that the introgressed lines continue to perform as well as the repeated (original) parent, since A. tauschii may have wild traits that are not suitable for cultivation. RNA-Seq-based genotyping was performed to detect SNPs between the selected lines and their parental wheat cultivars, which successfully revealed the chromosomal regions that were transmitted from the parental synthetic wheat line to the selected lines. These results demonstrate that this technique is efficient for the introgression and identification of traits from A. tauschii in the D-genome.

RNA-Seq-based DNA markers can be used to map genic regions, but not non-genic repeated regions, in the genome. The two examples of this technique being used described above demonstrate that RNA-Seq-based DNA markers can provide sufficient resolution for genetic distance-based analysis of A. tauschii and synthetic hexaploid wheat.

RNA-Seq based DNA marker analysis of other Triticeae species

Wild species of the genera Triticum and Aegilops are useful genetic resources for wheat breeding, as they can be used for interspecific crosses among cultivated and wild wheat species (Fig. 1). Tetraploids and hexaploids occur not only in Triticum but also in Aegilops and are thought to be derived from intercrosses between diploid species of Triticum and Aegilops (Tsunewaki 2009 and Chen et al. 2020). However, reference genome sequences are not yet available for most of these species. To evaluate intragenic DNA polymorphisms within species and nucleotide substitutions between species, we used RNA-Seq to analyze 12 species: T. urartu (AA genome), T. monococcum ssp. monococcum (AmAm genome), T. monococcum ssp. aegilopoides (AmAm genome) (Michikawa et al. 2019), A. umbellulata (UU genome) (Okada et al. 2018), A. speltoides (SS genome), A. bicornis (SbSb genome), A. searsii (SsSs genome), A. longissima (SlSl genome), A. sharonensis (SlSl genome) (Miki et al. 2019), A. caudata (CC genome), A. comosa (MM genome), and A. uniaristata (NN genome) (Tanaka et al. 2020). RNA-Seq of these diploid species allowed us to detect genome-wide SNPs and indels across chromosomes. These studies demonstrate the usefulness of RNA-Seq for detecting nucleotide polymorphisms, generating markers that distinguish each genome, and exploring the molecular evolution in these species. A summary diagram is shown in Fig. 1b.

A. umbellulata (UU genome) can be crossed with tetraploid wheat (T. turgidum AABB genome), which allows synthetic hexaploids (AABBUU genome) to be generated via ABU triploids. We performed RNA-Seq analysis of 12 representative accessions of A. umbellulata and reconstructed the transcripts of reads for each accession via de novo assembly (Okada et al. 2018). We anchored the deduced transcripts to the pseudomolecules of A. tauschii and barley, both of which are regarded as virtual chromosomes of A. umbellulata, and determined the distribution of SNPs and indels across the entire chromosomes. Genetic diversity in A. umbellulata was high despite its narrow habitat. No clear lineages were differentiated, and lower-frequency alleles were predominantly detected in A. umbellulata. These rare alleles might be the main source of the high genetic diversity of A. umbellulata.

Two wild diploid wheat species, Triticum monococcum ssp. aegilopoides and T. urartu, are closely related and harbor the Am and A genomes, respectively. T. urartu is the A-genome donor of tetraploid and common wheat, and T. monococcum ssp. monococcum is the cultivated form derived from the wild Am genome wheat subspecies aegilopoides. Since the Am and A genomes are genetically close, identifying a large number of markers that can discriminate between these genomes is challenging. Michikawa et al. (2019) detected genome-wide SNPs and indels from RNA-Seq data from the leaf transcripts of 15 accessions of these two diploid wheat species. The SNPs between the Am and A genomes, which were detected using the A-genome of common wheat as the reference genome, covered all of the chromosomes, facilitating the construction of PCR-based cleaved amplified polymorphic sequence (CAPS) markers that discriminate between the Am and A genomes. These markers effectively confirmed the addition of aegilopoides chromosomes to tetraploid wheat in nascent allohexaploid lines with AABBAmAm genomes. In addition, the markers allowed linkage maps to be constructed for mapping populations of aegilopoides accessions.

The efficient generation of markers in several Triticeae species prompted us to analyze genome differentiation in diploid wild wheat by applying RNA-Seq to wild wheat species with more diverse diploid genomes. We estimated genome differentiation based on an RNA-Seq-based survey of genome-wide polymorphisms throughout homoeologous loci in Triticum and Aegilops (Tanaka et al. 2020). The genome nomenclatures in these species were defined based on their affinity for chromosomal pairing. However, few studies have evaluated genome differentiation based on genome-wide nucleotide variations, especially in the three genomes of the genus Aegilops: A. caudata L. (CC genome), A. comosa Sibth. et Sm. (MM genome), and A. uniaristata Vis. (NN genome). Genetic divergence of the exon regions throughout all of the chromosomes was larger between the M and N genomes vs. the A and Am genomes. A. caudata had the second highest genetic diversity after A. speltoides, the putative B-genome donor of common wheat. In phylogenetic trees derived from nuclear and chloroplast genome-wide polymorphism data, the species with C, D, M, N, U, and S genomes were connected by short internal branches, suggesting that these diploid species emerged during a relatively short evolutionary period. The highly consistent nuclear and chloroplast phylogenetic topologies indicate that the nuclear and chloroplast genomes of the diploid Triticum and Aegilops species coevolved after their diversification into each genome, accounting for most of the genome differentiation among the diploid species. RNA-Seq-based analyses successfully revealed genome differentiation among the diploid Triticum and Aegilops species and supported the chromosome-pairing-based genome nomenclature system, except for the position of A. speltoides. Phylogenomic and epigenetic analyses of the intergenic and centromeric regions are needed to help clarify the basis of this inconsistency.

The S-genome of five species of section Sitopsis of the genus Aegilops is considered to be the origin of the B-genome in cultivated tetraploid and hexaploid wheat species, although the actual donor is unclear. Miki et al. (2019) attempted to elucidate the phylogenetic relationships among Sitopsis species by performing RNA-Seq of the coding regions of each chromosome. They extensively analyzed genome-wide polymorphisms in 19 accessions of the Sitopsis species in reference to the tetraploid and hexaploid wheat B genome sequences. Consequently, these polymorphisms were efficiently anchored to the B genome chromosomes. The results of genome-wide exon sequencing and subsequent phylogenetic analysis indicate that A. speltoides is likely the direct donor of all chromosomes of the wheat B genome. Only three chromosomal regions contradicted this phylogenetic relationship, and these exceptions could be explained by the higher recombination rates in distal regions of wheat chromosomes. The rate of genome differentiation during wheat allopolyploidization from S to B is not constant but varies along the chromosomes; recombination could affect this differentiation rate. This observation could potentially be generalized to genome differentiation during allopolyploidization in other plants.

As part of an IWGSC project, RNA-Seq-derived transcript sequences were used to generate BAC contigs for genome sequencing in wheat. Kobayashi et al. (2015) assembled 689 informative BAC contigs (hereafter referred to as contigs) representing 91% of the entire physical length of wheat chromosome 6B. The authors integrated the contigs into a radiation hybrid (RH) map of chromosome 6B, with one linkage group consisting of 448 loci with 653 markers. They then determined the order and direction of 480 contigs, corresponding to 87% of the total length of chromosome 6B. The authors also characterized contigs that contained part of the nucleolus organizer region or the centromere based on their positions on the RH map and the assembled BAC clone sequences. Analysis of the virtual gene order along chromosome 6B using information collected for the integrated map revealed the presence of several chromosomal rearrangements, representing evolutionary events that occurred on chromosome 6B. The physical map provided a high-quality, map-based reference sequence (IWGSC 2018).

As described above, RNA-Seq-based DNA marker generation could be performed for most diploid wild wheat species of Triticum and Aegilops. These species share orthologous exon sequences, allowing the sequences among species to be compared to estimate their evolutionary relationships. The finding that A. speltoides is likely the B genome donor of common wheat represents one of the most significant contributions of RNA-Seq-based DNA marker generation in wild wheat species. The use of markers for genome sequencing of wheat chromosome 6B is another important byproduct of RNA-Seq-based DNA marker generation.

Current status and future prospects for RNA-Seq-based DNA marker analysis in Triticeae

In conclusion, we reviewed information about (1) the establishment and application of an RNA-Seq-based DNA marker generation system for A. tauschii and (2) the use of RNA-Seq-based DNA markers for other diploid wild wheat species for molecular evolutionary studies.

Due to technical advancements in the ability to assemble large plant genomes (e.g., Triticeae species), chromosome-scale assembly can now be performed quickly and at a reasonable cost (Monat et al. 2019). Even a polyploid genome can be assembled using well-organized whole-genome shotgun reads and scaffold ordering techniques (e.g., Hi-C). These techniques can be used to reveal the variation across genome structures in a crop and related species in pan-genome studies (Sato 2020; Jayakodi et al. 2021). However, balancing economic importance and sequencing costs for the analysis of wild wheat species is an ongoing challenge. RNA-Seq has provided a valuable source of information for identifying polymorphisms in wild wheat accessions for use in genetic and evolutionary studies (Glémin et al. 2019; Miki et al. 2019; Tanaka et al. 2020), even when reference genomes for these species were not available and A. tauschii or barley diploid genomes were used to determine the genomes’ physical structures. As sequencing of each species in Triticeae is currently underway in the research community, the eventual availability of reference genomes for wild wheat species will further increase the efficiency of RNA-Seq-based DNA marker generation. cDNA sequencing has also been improved due to the development of single-molecule sequencing techniques provided by the PacBio and Nanopore platforms. If the gene models of each wild wheat genome are established with high-quality chromosome-scale assembly and mature transcript sequences, it might be possible to map RNA-Seq data onto the reference genome more efficiently.

The main advantage of using RNA-Seq to analyze multiple accessions is that the sequences are generated de novo, in contrast to analyses using pre-fixed polymorphism detection systems such as microarrays. Thus, RNA-Seq analysis has the flexibility to use reference sequences even if they are not from the same species as the reads. As a platform for analyzing the expressed portion of the genome and exons, RNA-Seq is a rapid and cost-effective technique that is particularly useful for analyzing the large genomes of Triticeae species.