ABSTRACT
Vitis vinifera cv. Cabernet Sauvignon is one of the world’s most widely cultivated red wine grape varieties and often used as a model for studying transcriptional networks governing berry development and metabolism. Here, we applied single-molecule sequencing technology to reconstruct the transcriptome of Cabernet Sauvignon berries during ripening. We added an error-correction step to the standard Iso-Seq pipeline that included using Illumina RNAseq reads to recover lowly-expressed transcripts. From 672,635 full-length non-chimeric reads, we produced 170,860 transcripts capturing 13,402 genes of the Cabernet Sauvignon genome. Full-length transcripts refined approximately one third of the gene models predicted using several ab initio and evidence-based methods. The Iso-Seq information also helped identify 563 additional genes, 4,803 new alternative transcripts, and the 5’ and 3’ UTRs in the majority of predicted genes. Comparisons with the gene content of other grape cultivars identified 549 Cabernet Sauvignon-specific genes, including 65 genes differentially regulated during ripening. Some of these genes were potentially associated with the phenylpropanoid and flavonoid pathways, which may influence unique Cabernet Sauvignon berry attributes. Over 23% of the 36,687 annotated genes in Cabernet Sauvignon had two or more alternative isoforms, predominantly due to intron retention and alternative acceptor and donor sites. We profiled the expression of all isoforms using short read sequencing and identified 252 genes whose alternative transcripts showed different expression patterns during berry development.