Analysis and annotation of the hexaploid oat seed transcriptome

Background: Next generation sequencing provides new opportunities to explore transcriptomes. However, challenges remain for accurate differentiation of homoeoalleles and paralogs, particularly in polyploid organisms with no supporting genome sequence. In this study, RNA-Seq was employed to generate and characterize the first gene expression atlas for hexaploid oat. Results: The software packages Trinity and Oases were used to produce a transcript assembly from nearly 134 million 100-bp paired-end reads from developing oat seeds. Based on the quality-parameters employed, Oases assemblies were superior. The Oases 67-kmer assembly, denoted dnOST (de novo Oat Seed Transcriptome), is over 55 million nucleotides in length and the average transcript length is 1,043 nucleotides. The 74.8x sequencing depth was adequate to differentiate a large proportion of putative homoeoalleles and paralogs. To assess the robustness of dnOST, we successfully identified gene transcripts associated with the biosynthetic pathways of three compounds with health-promoting properties (avenanthramides, tocols, beta-glucans), and quantified their expression. Conclusions: To our knowledge, this study provides the first direct performance comparison between two major assemblers in a polyploid organism. The workflow we developed provides a useful guide for comparable analyses in other organisms. The transcript assembly developed here is a major advance. It expands the number of oat ESTs 3-fold, and constitutes the first comprehensive transcriptome study in oat. This resource will be a useful new tool both for analysis of genes relevant to nutritional enhancement of oat, and for improvement of this crop in general.