A pan-transcriptome analysis shows that disease resistance genes have undergone more selection pressure during barley domestication

Abstract

Background: It has become clear in recent years that many genes in a given species may not be found in a single genotype thus using sequences from a single genotype as reference may not be adequate for various applications.

Results: In this study we constructed a pan-transcriptome for barley by de novo assembling 288 sets of RNA-seq data from 32 cultivated barley genotypes and 31 wild barley genotypes. The pan-transcriptome consists of 756,632 transcripts with an average N50 length of 1240 bp. Of these, 289,697 (38.2%) were not found in the genome of the international reference genotype Morex. The novel transcripts are enriched with genes associated with responses to different stresses and stimuli. At the pan-transcriptome level, genotypes of wild barley have a higher proportion of disease resistance genes than cultivated ones.

Conclusions: We demonstrate that the use of the pan-transcriptome dramatically improved the efficiency in detecting variation in barley. Analysing the pan-transcriptome also found that, compared with those in other categories, disease resistance genes have gone through stronger selective pressures during domestication.

Keywords: Barley; Cultivated; Novel transcripts; Pan-transcriptome; Resistance; Wild.

Fig. 1

Enhanced efficiency of CWA (cultivated+wild-assembly) in detecting variations in comparison with the use of the international reference genotype Morex. a) Novel transcripts and CDS not detected in Morex; and b) difference in the numbers of SNVs detected by using either Morex HC + LC transcripts or CWA as the reference. The comparison was conducted using RNA-seq data from of 19 accessions from one study as the sequences likely have similar qualities

Fig. 2

The workflow for identification and annotation of novel transcripts from cultivated and wild barley genotypes

Fig. 3

Difference in annotated protein domains between cultivated and wild barley. a Difference in the numbers of novel CDS. b Difference between all predicted CDS. The Y-axis shows the percentage of CDS with a specific protein domain. Protein domains marked with * indicate significant difference between cultivated and wild barley

Fig. 4

Overview of genetic differentiation between cultivated and wild barley with measured F_ST. Physical locations of transcripts from CWA were based on gmap results with the Morex genome assembly. a Frequency distribution of F_ST value for each locus. The X-axis indicates the value of F_ST. The red dash vertical line indicates the threshold at 0.72 (the 95th percentile threshold). b-h Chromosomal distributions of transcripts with F_ST larger than 0.72 with a window size 10 Mb. The X-axis indicates the physical positions of each chromosome and Y-axis shows the number of transcripts. The grey boxes on Y-axis indicates the approximate location of the centromere for each of the chromosomes

Fig. 5

Functional annotation of genes under strong selective pressures (F_ST > 0.72). a GO enrichment analysis; (b) count of genes (X-axis) in top 20 enriched KEGG pathways excluding the two global/overview maps (metabolic pathways and biosynthesis of secondary metabolites pathway)