De novo transcriptome characterization of Vitis vinifera cv. Corvina unveils varietal diversity

Abstract

Background: Plants such as grapevine (Vitis spp.) display significant inter-cultivar genetic and phenotypic variation. The genetic components underlying phenotypic diversity in grapevine must be understood in order to disentangle genetic and environmental factors.

Results: We have shown that cDNA sequencing by RNA-seq is a robust approach for the characterization of varietal diversity between a local grapevine cultivar (Corvina) and the PN40024 reference genome. We detected 15,161 known genes including 9463 with novel splice isoforms, and identified 2321 potentially novel protein-coding genes in non-annotated or unassembled regions of the reference genome. We also discovered 180 apparent private genes in the Corvina genome which were missing from the reference genome.

Conclusions: The de novo assembly approach allowed a substantial amount of the Corvina transcriptome to be reconstructed, improving known gene annotations by robustly defining gene structures, annotating splice isoforms and detecting genes without annotations. The private genes we discovered are likely to be nonessential but could influence certain cultivar-specific characteristics. Therefore, the application of de novo transcriptome assembly should not be restricted to species lacking a reference genome because it can also improve existing reference genome annotations and identify novel, cultivar-specific genes.

Figure 1

Genome coverage and sequence variation. a) Read counts normalized to gene length and log transformed (base=10) respect to the position on the 19 grape chromosomes. b) Classification of bases covered (≥3X) by feature type. c) Classification of SNPs based on the PN42004 genome annotation. d) Classification of indels based on the PN42004 annotation. e) Number of genes containing potentially disruptive mutations, plotted in logarithmic scale.

Figure 2

Contigs classification. a) Classification of contigs mapping onto the genome based on a comparison with the reference annotation V1 of assembly 12X. b) Distribution of the number of contigs per gene after filtering contigs for expression relative to the major isoforms (FMI). c) Classification of contigs and respective loci based on genomic region classes.

Figure 3

Classification of contigs mapping in putative novel loci. Classification of contigs mapping in putative novel loci, based on the coding potential calculated by CPC and on comparison with NCBI nr plant proteins and Rfam databases. Potentially coding contigs were classified as full ORFs begin with a start codon and end with an in-frame stop codon, or as partial ORFs if one of these two features was missing. The following two categories include contigs with a negative coding potential but Blast hit against the NCBI NR protein database or contigs with similarity with sequences in the Rfam RNA database.

Figure 4

Gene ontology (GO) classification of novel loci. Classification based on GO terms of 486 out of 2249 potentially novel protein-coding genes associated to least one GO term (level >1). a) Number of assignments to biological process GO onthology terms. b) Number of assignments to molecular function GO onthology terms.

Figure 5

Classification of contigs not mapping onto the reference genome. a) Distribution of unmapped contigs based on similarity to sequences in the NCBI non-redundant protein database and nucleic acid sequence databases (VvGI 8.0 and PN40024 raw reads). b) Distribution of contaminant sequences across different taxa.

Figure 6

Gene ontology (GO) classification of private genes. Assignments to GO terms of 100 out of 180 contig clusters corresponding to putative private genes. a) Number of assignments to biological process onthology terms. b) Number of assignments to molecular function onthology terms.

Figure 7

Expression profiles of 13,886 genes differentially expressed during berry development and withering. a) The differentially expressed genes were divided into four groups according to the expression profile: 1) repressed genes; 2) transiently repressed genes; 3) transiently induced genes; 4) induced genes. Black: novel gene loci in un-annotated regions of the genome; blue: putative private genes; red: average profile of the expression group. b) Number of assignments to GO Slim plant terms for the novel or private genes differentially expressed among at least two stages.