Sequencing 101: Isoforms

How does a genome of roughly 20,000 protein-coding genes¹ build a brain, an immune system, and hundreds of distinct cell types? Gene count alone falls well short of explaining it. Much of the answer lies in how each gene’s message is edited before it is ever used. A single gene can give rise to several distinct RNA messages, known as isoforms, and each one can carry its own instructions for building a protein. Most human genes produce more than one isoform², and a shift in which version is expressed can change protein function and contribute to disease even when overall gene expression appears unchanged. For researchers moving from gene-level RNA sequencing toward transcript-level questions, isoforms are often where the more revealing answers lie.

If you’re not sure whether isoforms make the difference in your research, our new guide highlights recent studies that show how full-length transcripts capture what short reads miss, and walks through sample planning, library preparation, and analysis in a practical seven-step checklist.

Explore the guide

What is an RNA isoform?

Think of a gene as the raw footage for a film. Before anything reaches the screen, editors cut away scenes that will not be used and splice the remaining shots together into a finished story. The same footage can become a theatrical release, a director’s cut, or a short trailer, and each version tells the story a little differently.

Inside the nucleus, a similar edit takes place. When a gene is transcribed, the first RNA copy, called pre-mRNA, contains exons, which typically carry coding information, along with introns that end up on the cutting-room floor. A large molecular complex called the spliceosome removes the introns and joins the exons together³, but the final cut is not fixed. Depending on cell type, developmental stage, or environmental signals, the splicing machinery may skip an exon, keep an intron, or use an alternative splice site within an exon. Transcripts can also open and close at different points, with alternative first exons and polyadenylation sites that reshape the untranslated regions (UTRs) controlling RNA stability, localization, and translation.

Each distinct cut of a gene’s final RNA message is an isoform. Some differ by only a few bases at a splice junction, while others encode proteins with different domains or even opposing functions.

Why alternative splicing matters for biology

Because isoforms rather than genes often determine functional outcomes, isoform-level resolution helps explain how a fixed genome supports so much biological variety. Splicing patterns shift across tissues, developmental stages, and environmental conditions, shaping processes that range from neuronal maturation to how crops respond to drought.

Gene-level measurements can obscure this activity. When two isoforms of the same gene move in opposite directions, their changes may cancel out in a gene count. Reference annotations also remain incomplete, particularly for less-studied tissues and non-model organisms.

Cancer research offers clear examples of where isoforms can make the difference. In lung adenocarcinoma, Englander and colleagues⁴ identified more than 180,000 full-length mRNA isoforms, 54% of them novel, including altered protein isoforms produced when an intron is retained in STAT2, an immune signaling gene. The level of STAT2 intron retention predicted how patients responded to checkpoint inhibitor immunotherapy, with greater retention of one intron associated with better survival. In gastric cancer, Huang and colleagues⁵ found two novel isoforms of ARID1A that begin at an alternative promoter and produce a truncated protein. These novel isoforms predicted progression-free survival more effectively than isoforms using the known promoter.

How long-read sequencing resolves full-length transcripts

Conventional short-read RNA sequencing breaks transcripts into short fragments and relies on software to reassemble them. To return to the film analogy, this is a bit like reviewing a pile of individual clips and trying to work out which version of the movie actually ended up in theaters. The approach measures gene expression well, but fragments rarely show which exons were joined together, so isoforms must be inferred and can end up incomplete or misassembled.

Highly accurate long-read sequencing, however, captures each polyadenylated transcript from its 5′ end to its 3′ end, including the poly(A) tail, in a single molecule with no assembly required, though read length is only part of the picture. Distinguishing closely related isoforms depends on placing each splice junction at the correct base, and reliable open reading frame prediction depends on reads free of the small insertion and deletion errors that can shift a predicted frame. High accuracy HiFi sequencing also allows sequence variants and isoform structure to be read together on the same molecule, supporting allele-specific isoform analysis.

Zajac and colleagues⁶ explored how these differences in technology play out by sequencing the same single-cell cDNA library from patient-derived kidney cancer organoids with both short and long reads. Although the two datasets overlapped substantially, the authors noted that short-read sequencing “does not provide isoform-level information or the information on structural variants, which can have a profound impact on deciphering cell-specific heterogeneity … In consequence, the community has turned to long-read sequencing that allows for full-length transcript sequencing.”

What recent publications reveal about isoforms

Joglekar and colleagues⁷ built a single-cell atlas of full-length isoforms across mouse and human brain cell types, developmental stages, and anatomical regions. They found that 72% of genes express different isoforms depending on cell type or context, and that isoform usage shifts dynamically as neural circuits mature.

In Alzheimer’s disease research, Gallo and colleagues⁸ used full-length isoform sequencing to profile APOER2, a receptor involved in APOE signaling, in postmortem brain tissue from people with Alzheimer’s disease and matched controls. They identified more than 200 unique isoforms, including some specific to brain region or disease state. In neurons, one Alzheimer’s-associated isoform altered receptor processing and reduced synapse numbers by roughly a third compared with a control-associated isoform.

Splicing also plays a central role in rare disease, where an estimated 15% or more of pathogenic mutations are thought to act by disrupting splicing⁹. Jaramillo Oquendo, Ferraro, and colleagues¹⁰ evaluated HiFi long-read RNA sequencing as a way to explain rare disease cases and to provide functional evidence for variants of unknown significance (VUSs). Twenty-four cases with VUSs from previous short-read DNA and RNA testing underwent long-read RNA sequencing following Kinnex protocols. Long reads confirmed 21 known aberrant splicing events and, in 8 cases, identified “additional effects on transcripts, enhancing interpretation of variants and either helping resolve pathogenicity or providing additional biological insights into the variant’s effects.” At least 10% of detected transcripts were novel and absent from existing catalogs, which the authors note may have “significant implications when standard annotations are used to quantify gene expression.”

These same patterns extend beyond human cells. As a widely used vertebrate model, the zebrafish offers insight into the molecular basis of human development and disease, yet much of its gene catalog was built from short reads. Bo and colleagues¹¹ generated a full-length transcriptome of zebrafish embryos spanning 21 stages of embryogenesis and identified 2,113 previously unannotated genes along with 33,018 novel isoforms of known genes. Many showed stage-specific expression and splicing during key developmental transitions. According to the authors, “The PacBio full-length transcriptome greatly improves the gene annotation of the zebrafish,” offering a more complete framework for studying gene regulation, development, and disease.

Taking the next step in transcript research

Isoforms connect genetic variation to protein function in ways gene-level data often miss, and a full-length view can refine variant interpretation and reveal disease mechanisms. For labs ready to bring isoform resolution into their own research, practical guidance on study design, sample requirements, and analysis available in the isoform guide can make the transition smoother.

Plan an isoform study

References

1. Amaral, P., et al. (2023). The status of the human gene catalogue. Nature, 622(7981), 41-47. https://doi.org/10.1038/s41586-023-06490-x
2. Wang, E. T., et al. (2008). Alternative isoform regulation in human tissue transcriptomes. Nature, 456(7221), 470-476. https://doi.org/10.1038/nature07509
3. Will, C. L., & Lührmann, R. (2011). Spliceosome structure and function. Cold Spring Harbor Perspectives in Biology, 3(7), a003707. https://doi.org/10.1101/cshperspect.a003707
4. Englander, R. P., et al. (2025). Intron retention regulates STAT2 function and predicts immunotherapy response in lung cancer. bioRxiv. https://doi.org/10.1101/2025.08.19.671121
5. Huang, K. K., et al. (2021). Long-read transcriptome sequencing reveals abundant promoter diversity in distinct molecular subtypes of gastric cancer. Genome Biology, 22(1), 44. https://doi.org/10.1186/s13059-021-02261-x
6. Zajac, N., et al. (2025). Comparison of single-cell long-read and short-read transcriptome sequencing via cDNA molecule matching: quality evaluation of the MAS-ISO-seq approach. NAR Genomics and Bioinformatics, 7(3), lqaf089. https://doi.org/10.1093/nargab/lqaf089
7. Joglekar, A., et al. (2024). Single-cell long-read sequencing-based mapping reveals specialized splicing patterns in developing and adult mouse and human brain. Nature Neuroscience, 27(6), 1051-1063. https://doi.org/10.1038/s41593-024-01616-4
8. Gallo, C. M., Kistler, S. A., Natrakul, A., Labadorf, A. T., Beffert, U., & Ho, A. (2024). APOER2 splicing repertoire in Alzheimer’s disease: Insights from long-read RNA sequencing. PLoS Genetics, 20(7), e1011348. https://doi.org/10.1371/journal.pgen.1011348
9. Baralle, D., Lucassen, A., & Buratti, E. (2009). Missed threads. The impact of pre-mRNA splicing defects on clinical practice. EMBO Reports, 10(8), 810. https://doi.org/10.1038/embor.2009.170
10. Jaramillo Oquendo, C., Ferraro, F., et al. (2026). HiFi long-read RNA sequencing enhances clinical diagnostics in rare disorders. European Journal of Human Genetics, 34(6), 840-851. https://doi.org/10.1038/s41431-026-02042-9
11. Bo, J., Fang, W., Wang, J., He, S., & Yang, L. (2025). High resolution of full-length RNA sequencing deciphers massive transcriptome complexity during zebrafish embryogenesis. BMC Biology, 23(1), 155. https://doi.org/10.1186/s12915-025-02271-2