Abstract
Hairy vetch (Vicia villosa Roth) and smooth vetch (V. villosa Roth var. glabrescens) are important cover crops and legume forage with great economic and ecological values. Due to the large and highly heterozygous genome, full-length transcriptome reconstruction is a cost-effective route to mining their genetic resources. In this study, a hybrid sequencing approach combining SMRT and NGS technologies was applied. The results showed that 28,747 and 40,600 high-quality non-redundant transcripts with an average length of 1808 bp and 1768 bp were generated from hairy vetch and smooth vetch, including 24,864 and 35,035 open reading frames (ORFs), respectively. More than 96% of transcripts were annotated to the public databases, and around 25% of isoforms underwent alternative splicing (AS) events. In addition, 987 and 1587 high-confidence lncRNAs were identified in two vetches. Interestingly, smooth vetch contains more specific transcripts and orthologous clusters than hairy vetch, revealing intraspecific transcript diversity. The phylogeny revealed that they were clustered together and closely related to the genus Pisum. Furthermore, the estimation of Ka/Ks ratios showed that purifying selection was the predominant force. A putative 3-dehydroquinate dehydratase/shikimate dehydrogenase (DHD/SDH) gene underwent strong positive selection and might regulate phenotypic differences between hairy vetch and smooth vetch. Overall, our study provides a vital characterization of two full-length transcriptomes in Vicia villosa, which will be valuable for their molecular research and breeding.