High-Throughput 3’ End Sequencing Reveals Novel Features in the Tomato Genome

Although modern genome sequencing methods manage to decrypt increasingly complex genomes, gene model prediction and genome annotation remain a difficult challenge. In this study, we compared an RNA 3’ end sequencing approach (TRANSeq) and a full-length transcript sequencing method (TruSeq, Illumina) to determine the gene expression profile of ripening tomato fruits. We found that both methods detect gene expression similarly, but notably TRANSeq could discriminate more efficiently between highly similar members of gene families, providing a more precise transcriptome profile. Our finding demonstrates that in each gene family, a single member displays high expression levels, while the other members exhibit low expression levels, suggesting the functional divergence of gene duplicates. In addition, we combined these two RNA-seq methods to improve the annotation of existing tomato gene models, and re-annotated >45% of tomato genes at their 3’end. These refined gene models will be incorporated into the next draft of the tomato genome (ITAG3.0; www.solgenomics.net). These findings suggest the importance of novel sequencing methods to assist genome annotation. We therefore expect that in the future 3’ end sequencing methods will be used to facilitate gene model predictions in all organisms. Moreover, these methods strongly reduce the sequencing cost per RNA sample, and in light or the results presented here, we suggest they become the methods of choice for transcriptomic analyzes.