Advancing Functional Transcriptomics in Zebrafish with High-accuracy Full-length RNA Sequencing

Zebrafish (Danio rerio) is a powerful vertebrate model organism with strong genetic and physiological similarity to humans, yet its use in large-scale transcriptomic research remains constrained by incomplete gene annotations, inefficient ribodepletion methods, and limited transcript-level resolution. To tackle these challenges, we applied CapTrap-seq, a platform-agnostic long-read RNA sequencing approach combining cap-trapping with oligo(dT) priming to selectively capture 5'-capped, full-length transcripts, to zebrafish developmental stages and adult tissues. We further introduce a size-selection step that substantially improves recovery of longer RNA molecules without compromising quantitative accuracy. Benchmarking against the template-switching oligo (TSO) approach demonstrated that CapTrap-seq enables accurate and reproducible transcript reconstruction in a non-mammalian system without requiring external ribodepletion or validation resources. Comparative analysis across multiple long-read catalogues showed that CapTrap-seq detected the largest number of biologically and clinically relevant genes, including oxidative phosphorylation, cardiac, and Online Mendelian Inheritance in Man (OMIM) disease genes, while revealing extensive isoform diversity absent from current annotations. Analysis of the carmn and dancr lncRNA loci further demonstrated the ability to resolve complex splicing landscapes, uncovering novel full-length isoforms with distinct domain architectures not represented in existing zebrafish reference annotations. CapTrap-seq thus emerges as a robust, genome-agnostic framework for high-quality transcriptome characterisation in zebrafish and other under-annotated species, with broad implications for functional genomics and translational research.