LsRTDv1, a reference transcript dataset for accurate transcript‐specific expression analysis in lettuce

SUMMARY Accurate quantification of gene and transcript‐specific expression, with the underlying knowledge of precise transcript isoforms, is crucial to understanding many biological processes. Analysis of RNA sequencing data has benefited from the development of alignment‐free algorithms which enhance the precision and speed of expression analysis. However, such algorithms require a reference transcriptome. Here we generate a reference transcript dataset (LsRTDv1) for lettuce (cv. Saladin), combining long‐ and short‐read sequencing with publicly available transcriptome annotations, and filtering to keep only transcripts with high‐confidence splice junctions and transcriptional start and end sites. LsRTDv1 identifies novel genes (mostly long non‐coding RNAs) and increases the number of transcript isoforms per gene in the lettuce genome from 1.4 to 2.7. We show that LsRTDv1 significantly increases the mapping rate of RNA‐seq data from a lettuce time‐series experiment (mock‐ and Botrytis cinerea‐inoculated) and enables detection of genes that are differentially alternatively spliced in response to infection as well as transcript‐specific expression changes. LsRTDv1 is a valuable resource for investigation of transcriptional and alternative splicing regulation in lettuce. Significance Statement Integrating long‐ and short‐read sequencing data from multiple lettuce tissues and treatments with existing publicly available transcriptome annotations led to the development of LsRTDv1. This is a reference transcript dataset that identifies novel genes and significantly increases the number of transcript isoforms per gene providing a valuable resource for accurate quantification of transcript‐level expression, and hence both transcriptional and post‐transcriptional regulation, in this leafy vegetable crop.