MicroExonator enables systematic discovery and quantification of microexons across mouse embryonic development

• Main Authors: Guillermo E. Parada, Roberto Munita, Ilias Georgakopoulos-Soares, Hugo J. R. Fernandes, Veronika R. Kedlian, Emmanouil Metzakopian, Maria Estela Andres, Eric A. Miska, Martin Hemberg
• Format: Article
• Language: English
• Published: BMC 2021-01-01
• Series: Genome Biology
• Subjects: Microexons; Splicing; Alternative splicing; Neuronal development; Single-cell; Reproducible software
• Online Access: https://doi.org/10.1186/s13059-020-02246-2

Abstract Background Microexons, exons that are ≤ 30 nucleotides, are a highly conserved and dynamically regulated set of cassette exons. They have key roles in nervous system development and function, as evidenced by recent results demonstrating the impact of microexons on behaviour and cognition. However, microexons are often overlooked due to the difficulty of detecting them using standard RNA-seq aligners. Results Here, we present MicroExonator, a novel pipeline for reproducible de novo discovery and quantification of microexons. We process 289 RNA-seq datasets from eighteen mouse tissues corresponding to nine embryonic and postnatal stages, providing the most comprehensive survey of microexons available for mice. We detect 2984 microexons, 332 of which are differentially spliced throughout mouse embryonic brain development, including 29 that are not present in mouse transcript annotation databases. Unsupervised clustering of microexons based on their inclusion patterns segregates brain tissues by developmental time, and further analysis suggests a key function for microexons in axon growth and synapse formation. Finally, we analyse single-cell RNA-seq data from the mouse visual cortex, and for the first time, we report differential inclusion between neuronal subpopulations, suggesting that some microexons could be cell type-specific. Conclusions MicroExonator facilitates the investigation of microexons in transcriptome studies, particularly when analysing large volumes of data. As a proof of principle, we use MicroExonator to analyse a large collection of both mouse bulk and single-cell RNA-seq datasets. The analyses enabled the discovery of previously uncharacterized microexons, and our study provides a comprehensive microexon inclusion catalogue during mouse development.