Reviews
Genetic resources and pangenome analysis of barley
2025 Volume 75 Issue 1 Pages 13-20
Details
2025 Volume 75 Issue 1 Pages 13-20
Barley (Hordeum vulgare) is widely cultivated, ranking fourth in cultivation area among cereal crops worldwide. Many wild and cultivated barley accessions have been collected and preserved in crop genebanks throughout the world. Barley has a large genome (~5 Gbp) that has recently been sequenced and assembled at the chromosome level by the international research community. The community also is sequencing accessions representing the diversity of both domesticated and wild barley to provide genome-wide genotyping information for pangenome analysis. Given that the pangenome represents the universe of genome sequences existing in a species, the long-term goal of this project is to obtain high-quality genome sequences of the major barley accessions worldwide. As each accession is annotated, the capacity to explore structural differences is enhanced by the increased understanding of the diversity of the barley genome, which will facilitate efficient development of cultivars for human consumption. This review describes our current knowledge of barley genome diversity and proposes future directions for basic and applied research of the barley pangenome.
Barley (Hordeum vulgare L.) is a self-pollinating crop with a diploid genome consisting of seven chromosomes (2n = 2x = 14). The estimated genome size of barley is 5.10 Gbp (IBSC 2012), comprising over 80% repetitive elements. Among Poaceae species, the genetics of barley has been particularly well studied. In addition, several mutant traits have practical importance and can be used as model traits in cereal crops. Notably, many barley and wheat (Triticum aestivum) genes have similar functions. Therefore, information about a gene in barley can be readily applied to identify the genes responsible for a similar trait in wheat, as demonstrated by the genome editing of a triple-knockout mutant of the wheat gene Quantitative trait locus seed dormancy 1 (Qsd1) (Abe et al. 2019), which was originally identified in barley (Sato et al. 2016a).
The genus Hordeum comprises over 30 species, including cultivated barley (Hordeum vulgare ssp. vulgare) (Blattner 2009). Of the gene pools listed in Fig. 1, only the ancestral form of cultivated barley, H. vulgare ssp. spontaneum, belongs to the primary gene pool of cultivated barley. This ancestral form of wild barley shares the same genome as domesticated barley and is classified as a barley subspecies (Bothmer et al. 2003). Crosses of cultivated barley with H. vulgare ssp. spontaneum show no incompatibility barriers; hence, there is a full capacity for gene transfer.
Gene pools of cultivated barley (Hordeum vulgare ssp. vulgare) (Bothmer et al. 2003).
H. vulgare shares the basic genome H with H. bulbosum, which is considered to be the only species in the secondary gene pool (Fig. 1). H. bulbosum has been used for crossing with cultivated barley to produce hybrids or for haploid induction due to chromosome elimination of H. bulbosum in hybrid progeny. Other Hordeum species are less closely related to cultivated barley and are classified as the tertiary gene pool of cultivated barley (Fig. 1). Most of these species share variants of the basic genome I, whereas two widespread weeds, H. marinum and H. murinum, possess distinct genomes (Xa and Xu, respectively; Bothmer et al. 1995).
Before barley was domesticated, the ancestral wild form of barley (H. vulgare ssp. spontaneum) was used as a human food, possibly to make an early form of flatbread using grains milled with stone tools, and the dough was baked in an open fire. Both wild and domesticated barley have been found in archaeological sites in the Fertile Crescent, which is considered to be the origin of barley domestication, dating back ca. 10,000 years (Zohary et al. 2012). As described by Sato (2020), wild barley kernels can be distinguished from domesticated barley kernels by their brittle and smooth rachis, which contrasts with the non-brittle and tough rachis of domesticated barley (Pourkheirandish et al. 2015). Millions of generations of this wild barley provide a source of diversity to the present-day cultivated form, although domestication partially narrowed the diversity of barley. Soon after its domestication, mutations responsible for agronomically valuable traits, e.g., six-rowed spikes (Komatsuda et al. 2007) and hull-less caryopses (Taketa et al. 2008), have been selected.
Allele diversity at agriculturally important loci associated with diverse ecological conditions and the different uses of this cereal (human food, animal feed, and malt production) have made barley landraces well suited for cultivation throughout the world. Naturally occurring diversity in locally adapted landraces and cultivars developed from landraces remains the most common source of available diversity for the crossbreeding of barley (Fischbeck 2003). Beyond crossbreeding, research on induced mutations to improve barley was initiated in the 1920s. Great efforts were made to increase the frequency of useful mutations and to control the process (Lundqvist and Franckowiak 2003). The numerous barley mutants produced by these efforts have greatly facilitated basic research and have contributed to practical breeding programs.
Barley germplasm has been extensively collected owing to the economic importance of this crop for agricultural production and for malting and producing animal feed (Ullrich 2011). Thus, most germplasm collections are composed of landraces of Hordeum vulgare ssp. vulgare. According to Knüpffer (2009), there are over 485,000 accessions of Hordeum residing in more than 200 genebanks worldwide. These collections include 299,165 accessions of H. vulgare ssp. vulgare, 32,385 accessions of H. vulgare ssp. spontaneum, and 4,681 accessions of other wild Hordeum species. From these accessions, an international barley core collection consisting of approximately 1,500 accessions was developed (Knüpffer and Hintum 2003).
Genotyping entire barley genebank collections generates comprehensive information about the genetic diversity and population structure within the accessions and provides efficient tools for barley breeding. The long-term goal is to manage and maintain collections of plant genetic resources by integrating biological and digital sequence information to enhance crop diversity and breeding. The first step is to collect genome-wide genotyping data for all accessions in a genebank collection. Several platforms can be used for genome-wide genotyping of barley. The Illumina iSelect single nucleotide polymorphism (SNP) array (Bayer et al. 2017) is an easy choice, but it may introduce ascertainment bias (Moragues et al. 2010) due to the pre-selected germplasm used to identify SNPs. Exome capture sequencing and whole-genome shotgun analysis require the construction of libraries, which prohibits the analysis of a large number of samples using complexity reduction and parallel sample processing.
Researchers at the Leibniz Institute of Plant Genetics and Crop Plant Research (IPK) in Gatersleben, Germany used genotyping-by-sequencing (GBS) to collect genotypic data for all ~22,000 barley accessions collected in the German federal genebank (Milner et al. 2019). GBS involves digesting DNA with two different restriction enzymes (e.g., PstI and MspI) prior to high-throughput sequencing (Poland et al. 2012). GBS does not require prior knowledge of target sequences or pre-analysis to design markers. GBS is a highly streamlined, multiplexed wet-lab protocol that relies only on standard laboratory techniques and reagents (Elshire et al. 2011). Although more powerful marker-based platforms may be available in the future, GBS is becoming a major platform for genotyping international barley collections in genebanks worldwide.
The International Barley Genome Sequencing Consortium (IBSC) was established in 2006 to generate a high-quality barley genome sequence using a combination of sequencing mapped gene-bearing BAC clones (BAC-by-BAC strategy) and whole-genome shotgun sequencing (IBSC 2012). From the 5.10 Gbp genome of American malting barley cultivar ‘Morex’, IBSC developed a physical map for 4.98 Gbp, with more than 3.90 Gbp anchored to a high-resolution genetic map, including 26,159 high-confidence genes with cDNA sequences and homology support from other plant genomes. More than 80% of the genome is occupied by repeat sequences.
Chromosome-scale assembly of barleyFor chromosome-scale, high-quality assembly of the barley genome, each BAC clone from the minimum tiling path was barcoded and sequenced using Illumina short reads. A high-resolution genetic map and a highly contiguous optical map (Irys system, BioNano Genomics Inc.) were then combined to construct super-scaffolds composed of merged assemblies from individual BACs. Finally, chromosome conformation capture sequencing (Hi-C) was used to order and orient the BAC-based super-scaffolds. The final chromosome-scale assembly represents 4.79 Gbp (~95%) of the genome of the ‘Morex’ reference assembly (MorexV1) (Mascher et al. 2017). Mapping of transcriptome data identified 39,734 high-confidence loci and 41,949 low-confidence loci based on sequence similarity to related species.
Assembly pipelineWhen the IBSC planned the BAC-by-BAC strategy for genome assembly (IBSC 2012), whole-genome shotgun assembly of the barley genome using Illumina short reads (150–300 bp) was performed as a supplementary technique, since less than half of the total genome was assembled owing to the failure of redundant repeat sequences to assemble properly (Sato et al. 2016b). However, chromosome-scale assembly using only short reads became available after the release of NRGene’s commercial genome assembly service DeNovoMAGIC.
The open-source assembly pipeline TRITEX (Monat et al. 2019) has also been developed. The purpose of the TRITEX pipeline is to construct chromosome-scale sequence assemblies for Triticeae genomes from different Illumina sequencing datasets within several weeks. The target applications of this technique include pangenomic studies in barley and wheat. The TRITEX pipeline only works with Illumina sequencing data of sufficient coverage and for a certain set of libraries. Paired-end and mate-pair datasets suitable for use with the pipeline are required, together with chromosome conformation capture sequencing (Hi-C and Dovetail Omni-C) data suitable for use with the pipeline and chromium data from 10X Genomics. The TRITEX pipeline must be run on a powerful Unix server with more than 1 TB of RAM. The pipeline requires a few weeks to assemble short-read data from barley and other plants with large genomes (such as wheat), which limits the number of accessions used for pangenome analysis.
Second version of the reference genome assembly of ‘Morex’The TRITEX pipeline was used to construct the second version of the reference genome assembly of ‘Morex’. This MorexV2 assembly resolves the problems of BAC-based reference sequences (Mascher et al. 2017) including (1) large sequence gaps, (2) redundancies, and (3) local mis-assemblies (Monat et al. 2019). The assembly metrics of the TRITEX MorexV2 assembly greatly exceed those of the BAC-by-BAC MorexV1 assembly (Fig. 2). The proportion of completely aligned full-length (FL)-cDNAs in TRITEX MorexV2 (Matsumoto et al. 2011) increased by ~9% compared to the BAC-by-BAC assembly.
Alignment of MorexV2 assembly scaffold_10x_163 to the MorexV1 assembly (Monat et al. 2019).
No single genome can represent the genetic diversity of a crop species. The concept of a ‘pangenome’ refers to the universe of genome sequences existing in a species (Jayakodi et al. 2021). Analysis of sequence variants segregating in multiple high-quality genome sequences is the goal of pangenome studies. Pangenomic methods have rapidly progressed over the past few years, and it is now practical to propose that common genomic analyses use a pangenome (Liao et al. 2023).
The starting point for pangenomics in barley was a comprehensive survey of species-wide diversity based on genome-wide genotyping of more than 22,000 barley accessions, mainly from the German federal genebank (Milner et al. 2019). To obtain a good representation of the primary barley gene pools, the authors selected accessions that were well separated by principal component analysis of GBS data on barley accessions (Fig. 3). In addition to these 18 gene pool representatives, these accessions included the reference genotype ‘Morex’ and one wild barley (H. vulgare ssp. spontaneum) accession from Israel ‘B1K-04-12’.
Nineteen domesticated barley genotypes representing the genetic diversity space, as revealed by principal component analysis (PCA) of genotyping-by-sequencing data from 19,778 domesticated barley varieties (Jayakodi et al. 2020).
The selected genotypes were assembled using the TRITEX assembly pipeline (n = 16), DeNovoMAGIC from NRGene (n = 3), or W2rap (https://doi.org/10.1101/110999) (n = 1). A comparison of the short-read assembly of ‘Morex’ to a long-read assembly of this genotype generated from PacBio long reads was performed to demonstrate that short-read assemblies are amenable to pangenomic analysis in barley. Although the assemblies of the 20 diverse accessions differed in terms of contiguity and the extent of gap sequences in the intergenic space, they had a similar representation of reference gene models (MorexV2) and were highly co-linear at the whole-chromosome scale (Jayakodi et al. 2020).
Annotation of the barley genome assembliesA similar proportion (~80%) of the assembled sequence of each genotype was composed of transposable elements (Jayakodi et al. 2020). De novo gene annotation using Illumina RNA-Seq and PacBio Iso-Seq data was performed for three genotypes: ‘Morex’, ‘Barke’, and the Ethiopian landrace ‘HOR 10350’. Gene models defined based on these three assemblies were consolidated and projected onto 17 other assemblies (gene projection). Between 35,859 and 40,044 gene models in each assembly were annotated by gene projection. The clustering of orthologous gene models yielded 40,176 orthologous groups. Of these, 21,992 occurred as a single copy in all 20 assemblies, 3,236 occurred in multiple copies in at least one of the 20 assemblies, 13,188 were absent from at least one assembly, and 1,760 were present in only one assembly (Jayakodi et al. 2020).
Structural variations in the 20 barley genome assembliesMultiple genome assemblies from representative accessions are a source of structural variants, which can be used to genotype other genebank accessions with low-coverage sequence data or genome-wide genotyping data. Using Assemblytics (Nattestad and Schatz 2016) software, Jayakodi et al. (2020) discovered presence–absence variation (PAV) by performing a pair-wise comparison of 19 chromosome-scale assemblies to the ‘Morex’ reference assembly. The authors identified 1,586,262 PAVs, ranging in size from 50 to 999,568 bp. PAV density was higher in distal, gene-rich regions with higher recombination rates.
Another method for estimating structural variations is to identify single-copy regions extracted from each of the 20 assemblies and cluster them into a non-redundant set of sequences (Jayakodi et al. 2020). The average cumulative size of a single-copy sequence in each accession was 478 Mb (ca. 9.5% of the genome assembly). The total size of non-redundant single-copy sequences was 638.6 Mb, including 402.5 Mb of single-copy sequences shared among all 20 genotypes and 235.9 Mb of variable sequences. On average, each of the 20 genotypes contained 2.9 Mb of single-copy sequences not present in any other assembly.
There are some chromosome inversions between these assemblies and MorexV2 identified using Hi-C data. These are confirmed as real inversions and not from mis-assemblies. Some of these inversions on the long arms of chromosomes are shown in Fig. 4.
Whole-genome alignments showing some examples of large chromosomal inversions identified using Hi-C data (Jayakodi et al. 2020).
Based on the analysis of single-copy sequences, Jayakodi et al. (2020) selected PAV markers in Morex. The black and red dots in the Manhattan plot in Fig. 5a denote single-copy sequences that are present and absent in Morex, respectively. To test the suitability of the single-copy pangenome for genetic analysis in a wider diversity panel without high-quality genome sequences, Jayakodi et al. (2020) collected whole-genome shotgun data (threefold coverage) for 200 domesticated and 100 wild accessions of barley. In genome-wide association scans, the pangenome marker that was most highly associated with lemma adherence covered the NUDUM (NUD) gene (Taketa et al. 2008) (Fig. 5a). All varieties of naked barley—in which lemmas can be easily separated from grains—are thought to trace back to a single mutational event, deleting the entire NUD sequence. The most highly associated PAV marker was a 16.7-kb region that is deleted in the naked accession ‘HOR 7552’ and that contains the NUD gene (Fig. 5b).
a, Genome-wide association scan for lemma adherence on the basis of present and absent variation (PAV) markers with Morex and other 19 pangenome assemblies. The black and red dots in the Manhattan plot denote single-copy sequences that are present and absent in Morex, respectively. b, The most highly associated PAV marker was a 16.7-kb region that is deleted in the naked accession HOR 7552 and that contains the NUD gene (Jayakodi et al. 2020).
Using a similar strategy to that of Jayakodi et al. (2020), two additional assemblies were established, including assemblies of wild accession ‘OUH602’ (Sato et al. 2021a) and barley cultivar ‘Haruna Nijo’ (Sakkour et al. 2022). These germplasms are key genetic and genomic resources at Okayama University that have been used as parents of mapping populations, genotypes for cDNA analysis, and BAC library development. The authors used the TRITEX pipeline and annotation methodologies from pangenome analysis for chromosome-scale assembly of the two genotypes. Assembly information and browsers are available at the NBRP barley website (http://earth.nig.ac.jp/~dclust/cgi-bin/index.cgi). The FL-cDNA sequences of ‘Haruna Nijo’ (Matsumoto et al. 2011, Sato et al. 2009) were mapped onto the assembly, which could serve as a key resource for gene identification in barley.
The recent development of long-read sequencing by circular consensus sequencing (CCS) on the PacBio platform alters the strategy of plant pangenome analysis. Mascher et al. (2021) compared the ability of current long-read sequencing platforms to rapidly generate contiguous sequence assemblies in barley (Fig. 6). Long-read assemblies are clearly superior to short-read assemblies in barley. A downsampling study indicated that 20-fold genome coverage of PacBio HiFi reads provide qualified assemblies in barley. A similar downsampling study in hexaploid wheat cultivar ‘Fielder’ by Sato et al. (2021b) indicated that 16-fold reads are sufficient for genome assembly in hexaploid wheat, but that 12-fold reads provide most of the information needed for genome assembly. Long reads can be used to construct accurate and complete assemblies of multiple accessions of a species to build pangenome infrastructures in Triticeae crops.
Alignments of MorexV3 and MorexV2 pseudomolecules in the terminal 10 Mb of the long arm of each chromosome. Gray lines indicate scaffold boundaries (Mascher et al. 2021).
The generation of gapless, telomere-to-telomere (T2T) sequence assemblies of plant chromosomes was first reported in Arabidopsis thaliana (Naish et al. 2021). T2T assemblies of plant genomes were subsequently reported in various plant species, including crops. Navrátilová et al. (2022) used the same strategy employed by plant T2T projects to analyze sequence gaps in the most updated reference genome sequence of MorexV3. They used long-read sequencing data for MorexV3 and ChIP-seq data for the centromeric histone variant CENH3 to estimate the abundance of centromeric DNA, ribosomal DNA, and subtelomeric repeats in the barley genome. However, they could not assemble the T2T sequence in barley since almost all centromeric sequences and 45S ribosomal DNA repeat arrays were absent from the MorexV3 pseudomolecules and that the majority of sequence gaps could be attributed to assembly breakdown in long stretches of satellite repeats.
There is also a gap between the estimated genome size of the ‘Morex’ draft genome (5.10 Gbp; IBSC 2012), the size (4.79 Gbp) of the MorexV1 assembly (Mascher et al. 2017), and the 4.0- to 4.5-Gbp genome sizes identified by pangenome analysis using the TRITEX pipeline (Jayakodi et al. 2020). The estimated genome sizes of the newer TRITEX MorexV2 assembly and the long-read MorexV3 assembly (both 4.2 Gbp) are not any longer than the initial estimates from the BAC-based physical map.
Pan-transcriptomesA pan-transcriptome describes the transcriptional and post-transcriptional consequences of genome diversity from multiple individuals within a species (Waugh et al. unpublished). In barley, deep sequencing of the transcriptome (RNA-seq) from ‘Morex’ and FL-cDNAs from ‘Haruna Nijo’ facilitated the annotation of the reference genome of ‘Morex’. The recent development of a single-molecule sequencing technique may also support the sequencing of long transcripts. An international collaborative project aimed at full-transcript sequencing of 20 pangenome genotypes is currently underway (Waugh et al. unpublished). These long-transcript sequences will also contribute to the annotation of each assembly in pangenome analysis. As it is essential to isolate intact mature transcripts (mRNA) to obtain FL-cDNA sequences, techniques are being developed to maintain transcript integrity.
Genomic information and browsersFor genome assembly and annotation, Ensembl Plants provides easy access to the most updated barley genome assembly (MorexV3), including the sequences of chromosomes, genes, transcripts, and predicted proteins. The same website also supports BLAST searches against the barley genome. Similarly, IPK allows users to conduct BLAST searches against all sequence resources published by the worldwide barley community, including activities related to the International Barley Sequencing Consortium. Pangenome graphs will be needed to replace the current single-genotype genome browser, as demonstrated for the human pangenome project. However, the resolution of sequence variation in barley is currently limited to 50 bp (Bayer et al. personal communication), and the use of multiple haplotype browsers is limited at present.
Genebank genomics forms the basis for the digital genebank system (Mascher et al. 2019). The ultimate goal of a digital genebank system is the de novo sequencing of all natural and engineered germplasm. Preparing a high-molecular-weight DNA sample for library construction for long-read sequencing runs currently takes a few weeks. The assembly (e.g., TRITEX) pipeline can assemble chromosome-scale sequences for a single barley accession within 3 to 4 weeks. A parallel sequencing and computational analysis strategy might save time, but it is not feasible to sequence thousands of accessions by high-quality chromosome-scale assembly using the current technical standards. An automatic gene annotation system must be developed, since the annotation process currently relies on slow, manual data curation steps. For these reasons, the combination of genotyping all genebank accessions and increasing the number of chromosome-scale assemblies of accessions representing barley diversity is a reasonable solution for the near future.
Contribution of pangenome information to breedingNatural and induced variation provides opportunities to analyze traits of interest. SNP haplotype–based characterization of genebank accessions and breeding germplasm is the current approach to standardize the materials used for plant breeding. High-quality genome sequences are essential for the basis of haplotype analyses and establishing worldwide digital genebank networks. Association genetics is a suitable way to utilize pangenomes to combine barley sequence variation with phenotypic variation via user-friendly pangenome browsers (Jayakodi et al. 2021).
Agronomically important traits are often controlled by multiple interacting genes, which demands deep knowledge of trait-based genetics. To better understand the contribution of genomic sequence variation to various traits, standardized phenotyping is also emerging as an essential goal. Systematic gene modification by genome editing may also provide a functional view of candidate genes of interest in multiple genotypes. Using recently updated sequencing techniques, it is also feasible to include secondary and tertiary gene pools of barley to explore the diversity of the genus Hordeum.
Finally, the combined information collected from genomic sequences and the systematic functional analysis of genes provides novel strategies for trait improvement via the sequence-based breeding of barley.
K.S. wrote the manuscript.
I thank the National Bioresource Project, Japan, for the barley resources stored at Okayama University described in this review. This work was supported by JSPS KAKENHI, Japan, grant number 23H00333.
