Distribution, functional impact, and origin mechanisms of copy number variation in the barley genome

Abstract

Background: There is growing evidence for the prevalence of copy number variation (CNV) and its role in phenotypic variation in many eukaryotic species. Here we use array comparative genomic hybridization to explore the extent of this type of structural variation in domesticated barley cultivars and wild barleys.

Results: A collection of 14 barley genotypes including eight cultivars and six wild barleys were used for comparative genomic hybridization. CNV affects 14.9% of all the sequences that were assessed. Higher levels of CNV diversity are present in the wild accessions relative to cultivated barley. CNVs are enriched near the ends of all chromosomes except 4H, which exhibits the lowest frequency of CNVs. CNV affects 9.5% of the coding sequences represented on the array and the genes affected by CNV are enriched for sequences annotated as disease-resistance proteins and protein kinases. Sequence-based comparisons of CNV between cultivars Barke and Morex provided evidence that DNA repair mechanisms of double-strand breaks via single-stranded annealing and synthesis-dependent strand annealing play an important role in the origin of CNV in barley.

Conclusions: We present the first catalog of CNVs in a diploid Triticeae species, which opens the door for future genome diversity research in a tribe that comprises the economically important cereal species wheat, barley, and rye. Our findings constitute a valuable resource for the identification of CNV affecting genes of agronomic importance. We also identify potential mechanisms that can generate variation in copy number in plant genomes.

Figure 1

Frequency spectrum of CNV. (A) Percentage of CNVs identified in one to 14 genotypes relative to the total number of events; (B) frequency spectra comparison between wild and cultivated barley.

Figure 2

Distribution of CNV per chromosome for all genotypes, wild barleys, and cultivated barleys. The bars represent percentages of CNVs assigned to each chromosome relative to the total number of contig fragments present on the corresponding chromosome. The single asterisk indicates that, considering all genotypes, the percentage of CNV on 4H is significantly lower compared to other chromosomes (t-test P value = 0.0002), while the double asterisk indicates the frequency of CNV on 4H in cultivated barley is significantly lower than wild barley (P value = 0.003 by t-test).

Figure 3

Distribution and frequency of structural variation across the seven barley chromosomes. The upper plots show, for each barley chromosome, all variants assigned to chromosome positions and the number of genotypes sharing each variant, with colors indicating the type of structural variation (blue=UpCNV; red=DownCNV/PAV; green=Up and Down; grey=no variation). The lower panels shown for each chromosome illustrate the proportions of copy number variants per 1.5M bp window with respect to the total number of fragments assigned to that window, with proportions represented by a color gradient from black (proportion =0) to yellow (proportion = 1).

Figure 4

Relationship between recombination rate and frequency of CNV. The black line represents the recombination trend calculated from the cM/Mb ratios along the physical map. All the chromosomes were combined and the window size was set to 10 Mb. The red dots represent the proportion of CNVs with respect to the total number of contig fragments in each 10 Mb bin.

Figure 5

Comparison between CNVs identified in wild and cultivated barley. (A) Venn diagram showing the overlap between regions affected by CNV in both subgroups. (B) Venn diagram illustrating the overlap in CNVs that affect coding sequences.

Figure 6

Examples of sequence alignments of contig fragments containing DownCNV/PAVs. The sequence of the barley cultivar Morex is shown at the top and the sequence of cultivar Barke at the bottom. (A) Schematic representation of how an insertion in Barke can lead to a DownCNV/PAV call. Sequence regions that are orthologous are connected by shaded areas. The additional sequence in Barke is depicted in light blue. The full contig fragment is composed of 10 overlapping probes. Those probes which overlap the breakpoint of the insertion will produce a low intensity signals or no signals, resulting in a reduced overall signal of the targeted contig fragment. (B) Contig fragments with multiple insertions/deletions. (C) Contig fragment with multiple deletions, including one that expands past the border of the fragment. (D) Contig fragment that contains additional sequences in Barke. (E) Contig fragment that contains an insertion/deletion that most likely originates from template slippage. The numbers in circles identify different types of insertions/deletions: 1, insertion/deletion that contains no obvious signature; 2, insertion/deletion that shows a typical signature of double-strand break repair via single-strand annealing (SSA); 3, insertion/deletion which contains filler sequence (indicated by a curly bracket) and that presumably is the result of DSB repair via synthesis-dependent strand annealing (SDSA); 4, insertion/deletion originated from template slippage of direct repeats (indicated by arrows).