| Summary: | Abstract Background The technological advances of RNA-seq and de novo transcriptome assembly have enabled genome annotation and transcriptome profiling in highly heterozygous species such as grapevine (Vitis vinifera L.). This work is an attempt to utilize a de novo-assembled transcriptome of the V. vinifera cultivar ‘Riesling’ to improve annotation of the grapevine reference genome sequence. Results Here we show that the transcriptome assembly of a single V. vinifera cultivar is insufficient for a complete genome annotation of the grapevine reference genome constructed from V. vinifera PN40024. Further, we provide evidence that the gene models we identified cannot be completely anchored to the previously published V. vinifera PN40024 gene models. In addition to these findings, we present a computational pipeline for the de novo identification of lncRNAs. Our results demonstrate that, in grapevine, lncRNAs are significantly different from protein coding transcripts in such metrics as length, GC-content, minimum free energy, and length-corrected minimum free energy. Conclusions In grapevine, high-level heterozygosity necessitates that transcriptome characterization be based on cultivar-specific reference genome sequences. Our results strengthen the hypothesis that lncRNAs have thermodynamically different properties than protein-coding RNAs. The analyses of both coding and non-coding RNAs will be instrumental in uncovering inter-cultivar variation in wild and cultivated grapevine species.
|
|---|