Differential transcript expression and differential transcript usage using Salmon and edgeR

Dong, Xueyi, Chen, Lizhong, Nie, Junli, Smyth, Gordon K., Chen, Yunshun
Nature Protocols · Springer Nature · 2026-11-16 · DOI: 10.1038/s41596-026-01443-w

Abstract

Differential transcript expression (DTE) and differential transcript usample information and experimentalsage (DTU) analyses provide high-resolution insights into transcript-level changes and regulatory mechanisms. Here we present a protocol built around the recently developed edgeR DTE and DTU frameworks, which incorporate quantification-uncertainty correction and new quasi-likelihood methods that have been demonstrated to outperform alternative tools in statistical power, false discovery rate control and computational efficiency. These methods are general and can be applied to any well-annotated reference transcriptome. To illustrate the workflow, we apply the protocol using the complete telomere-to-telomere (T2T-CHM13) human reference genome, whose improved assembly and transcript annotation offer enhanced accuracy for transcript-level analyses of human samples. The protocol uses Salmon for fast transcript quantification, incorporates strategies for organizing GFF-based annotations and employs edgeR for assessing DTE and DTU while accounting for read-to-transcript ambiguity and variation between biological replicates. Applying the workflow to an RNA-sequencing dataset from human lung adenocarcinoma cell lines, we show that the combination of Salmon and edgeR reliably identifies differentially expressed or differentially used transcripts between experimental conditions. The end-to-end protocol can be completed in under 2 h for quantification in Salmon and less than 10 min for statistical analysis in edgeR. It provides a flexible, high-performance framework for transcript-level analyses across diverse organisms and reference genomes and can be carried out by researchers with basic experience using the Unix command line and R. Using second-generation RNA-sequencing data as the input, the workflow begins with the preparation of a transcriptome index. A detailed procedure describes the subsequent transcript quantification and annotation, followed by the preprocessing of count data and model fitting. Differential transcript expression and differential transcript usage analyses are then conducted. The resulting data are plotted and transcript changes are interpreted from these plots. Studying changes in the abundance of a transcript and the proportional changes in transcript variants can help answer fundamental biological questions. This protocol describes a framework for differential transcript expression and usage analyses across diverse reference genomes.

Read at publisher →

Similar articles