A high quality Arabidopsis transcriptome for accurate transcript-level analysis of alternative splicing

Abstract

Alternative splicing generates multiple transcript and protein isoforms from the same gene and thus is important in gene expression regulation. To date, RNA-sequencing (RNA-seq) is the standard method for quantifying changes in alternative splicing on a genome-wide scale. Understanding the current limitations of RNA-seq is crucial for reliable analysis and the lack of high quality, comprehensive transcriptomes for most species, including model organisms such as Arabidopsis, is a major constraint in accurate quantification of transcript isoforms. To address this, we designed a novel pipeline with stringent filters and assembled a comprehensive Reference Transcript Dataset for Arabidopsis (AtRTD2) containing 82,190 non-redundant transcripts from 34 212 genes. Extensive experimental validation showed that AtRTD2 and its modified version, AtRTD2-QUASI, for use in Quantification of Alternatively Spliced Isoforms, outperform other available transcriptomes in RNA-seq analysis. This strategy can be implemented in other species to build a pipeline for transcript-level expression and alternative splicing analyses.

How does temperature affect splicing events? Isoform switching of splicing factors regulates splicing of LATE ELONGATED HYPOCOTYL (LHY)

“…Gene and transcript models for PTB1 , PTB2 , U2AF65A , and SUA reported in AtRTD2 (Zhang et al, ) are shown in Figure S2. The functional and the main non‐functional premature termination codon‐containing (PTC + ) PTB transcripts are referred to as fully spliced (FS) and alternative exon (AE), respectively.…”

Genome‐wide characterization of differential transcript usage in Arabidopsis thaliana

“…These were generated with the Flux simulator (Griebel et al, 2012), which takes as input the number of RNA molecules (molecule count) to simulate for each isoform, and can simulate single-or paired-end reads of different lengths. A compendium of 100 simulated samples was generated, containing molecule counts for all isoforms in the AtRTD2-QUASI A. thaliana reference transcriptome (Zhang et al, 2017b). AtRTD2-QUASI is a modified version of AtRTD2 that has been optimized for isoform expression quantification.…”

“…In November 2010, the TAIR10 annotation contained 27 416 protein coding genes, of which 21% had more than one known isoform (on average 2.37 isoforms per gene; Lamesch et al, 2012). In 2016, two new annotations were released: Araport11 in which 41% of 27 655 protein-coding genes had more than one isoform with on average 2.94 isoforms per gene (Cheng et al, 2017); and AtRTD2 in which 49% of 27 667 protein-coding genes (60% of 22 453 introncontaining) had more than one isoform with on average 4.44 isoforms per gene (Zhang et al, 2017b). However, the large number of isoforms per gene in AtRTD2 potentially complicates expression quantification as an increased number of isoforms negatively impacts quantification accuracy (Hayer et al, 2015;Kanitz et al, 2015;Zhang et al, 2017a).…”

“…Here we present a large-scale dataset containing isoform expression levels, isoform dominance calls and DTU calls for A. thaliana across 206 public RNA-Seq experiments, covering different combinations of stress and hormone treatments, and a wide array of vegetative and reproductive tissues. For this, the AtRTD2 reference transcriptome was used (Zhang et al, 2017b), which provides the most extensive and accurate collection of transcripts for A. thaliana to date, and consists of a merger between annotations from TAIR10 (Lamesch et al, 2012), Araport11 (Cheng et al, 2017), AtRTD (Zhang et al, 2015b) and a new transcript assembly based on 129 RNA-Seq libraries (Zhang et al, 2017b). Based on the generated dataset, we provide a genome-wide overview of DTU in relation to gene function by integrating sample metadata, Gene Ontology (GO) annotation and protein domain analysis.…”

Computational aspects underlying genome to phenome analysis in plants

“…Short‐read sequencing technologies, in conjunction with annotated reference genomes, can be readily applied to a variety of biological questions, including detection (Zhang et al ., ) and analysis of gene expression (Ezer et al ., ), DNA methylation (Zhong et al ., ), identification of transcription factor binding sites and the detection of causal regions and mutations in mutant screens (James et al ., ; Klap et al ., ) or populations (Thoen et al ., ). However, the importance of next‐generation sequencing beyond the context of a single reference accession has long been recognized (Varshney et al ., ).…”