Genetic investigation of seed formation during berry development using RNA-Seq reveals transcriptional changes in grapevine (Vitis vinifera L.)

With the recent advances in the next generation sequencing technology, RNA-Seq has become a popular and more comprehensive approach towards quantification of the transcriptome, detection of novel splice variants, SNPs and discovery of unknown exons or genes by direct ultra-high-throughput sequencing of cDNAs. Similarly the successful sequencing and release of Vitis vinifera genomic data has contributed to characterize biological function of genes involved in important agronomic traits in grapevine. Nowadays, quantitative traits such as disease resistance, larger fruit size and higher fruit number per plant, better flavour, taste, nutritional quality, reduced seed number and size are targets for manipulation in grape genetic improvement programmes. Considering table-grapes, seedlessness is one of the most appreciated quality traits, as breeding focus shifts towards combining seedlessness with other quality traits such as larger berry size, Muscat flavour or crispiness; we shed light on the transcriptional responses that are related to seed development in V. vinifera. Illumina mRNA-Seq technology was used to sequence six cDNA libraries prepared from total RNA collected at three key developmental stages, Pre-flowering, Full flowering and Full ripening, in a seeded widespread wine cultivar (wild-type) and a spontaneous seedless somatic variant (mutant). Over 100 million reads, between 17 and 225 bp in length from each library, were mapped to 12x Grape Genome assembly and V1_mRNA annotations to define gene structure and quantify transcript abundance. Genic distribution of uniquely mapped reads shows 76 percent coverage of V1_mRNA (annotated exons). We found 27,496 genes to be expressed during seed developmental process, 7119 of which were expressed in a stage specific manner. To estimate variability in our data, uniquely mapped read counts were subjected to differential expression analysis using the software DESeq in R programme. The sets of differentially expressed genes during seed development were compared between the wild-type and the mutant. To reveal functionally related genes, an in-house R script was written to group genes with similar expression pattern based on the P-adjusted values. Functional annotation and pathway analysis were also performed. Finally we validated the expression of selected putative candidate genes using qRT-PCR.