Unlocking isoform programs underlying brain development with long-read single-cell RNA sequencing | LC26

Abstract
Almost all human genes produce multiple mRNA products (RNA isoforms), many with distinct or even opposing functions. In the human brain, isoform diversity is exceptionally high and tightly linked to development and disease, yet isoform usage has remained largely invisible at the level of individual cell types, limiting our ability to connect gene regulation to neural function and pathology. To address this, we developed long-read single-cell RNA sequencing (LR scRNA-seq) methods and software that enable high-resolution profiling of known and novel isoforms at single-cell resolution. Here, we introduce and implement a faster and more accurate version of our widely used FLAMES analysis framework and apply it to study neurogenesis in differentiating human cortical organoids. LR scRNA-seq enabled high-resolution identification of cell types and subtypes, including distinct radial glial progenitors and excitatory neurons. Across the dataset, we detected more than 170,000 unique isoforms, including over 10,000 previously unannotated transcripts, 1,747 novel exon loci and 8,958 unique micro-exon loci. Micro-exon usage was highly cell-type-specific, with strong enrichment in excitatory neurons. Thousands of isoforms displayed differential expression linked to synaptic transmission, neuronal projection, axonogenesis, and neuronal maturation. Notably, genes such as PKM and GPM6A showed ubiquitous gene-level expression yet cell-type-specific isoform usage, demonstrating that gene-level measurements alone can obscure biologically meaningful regulatory variation. By resolving isoform expression across cell types and developmental trajectories, this work substantially expands the resolution at which transcriptomic regulation can be studied in the human brain and provides a framework for linking RNA isoforms to neurodevelopmental processes and disease.

Biography
Sefi (Yair) Prawer is a Postdoctoral Research Fellow at The University of Melbourne, specialising in nanopore long-read single-cell technologies. His research leverages these approaches to investigate isoform-level regulatory programs that drive human brain development and contribute to neuropsychiatric disease aetiology. He actively generates nanopore long-read single-cell datasets and develops analysis workflows, tutorials, and software tools to support data analysis and visualisation. His contributions include tools such as FLAMES for long-read single-cell analysis and BLAZE for long-read single-cell barcode detection.