Partitioning RNAs by length improves transcriptome reconstruction from short-read RNA-seq data

The accuracy of methods for assembling transcripts from short-read RNA sequencing data is limited by the lack of long-range information. Here we introduce Ladder-seq, an approach that separates transcripts according to their lengths prior to sequencing and uses the additional information to improve the quantification and assembly of transcripts. Using simulated data, we demonstrate that a kallisto algorithm extended to process Ladder-seq data quantifies transcripts of complex genes with substantially higher accuracy than conventional kallisto. For reference-based assembly, a modified StringTie2 algorithm reconstructs a single transcript with 30.8% higher precision than its conventional counterpart and is >30% more sensitive for complex genes. For de novo assembly, a modified Trinity algorithm correctly assembles 78% more transcripts than conventional Trinity, while improving precision by 78%. In experimental data, Ladder-seq reveals 40% more genes harboring isoform switches compared with conventional RNA-seq and unveils widespread changes in isoform usage upon m6A depletion by Mettl14 knock-out. Overall design: RNA sequencing of neural progenitor cells from Mettl14 wild type and conditional knock-out mice with four replicates per genotype. Prior to sequencing mRNA from each sample was separated by length into 7 distinct bands. Each band from each sample has a unique barcode.