Scientists at King’s College London, the Francis Crick Institute and international partners have identified hundreds of previously overlooked gene segments called microexons and shown that DNA variants affecting their regulation may explain autism risk in families where standard sequencing found no answer.
The work, published in Nature on 14 October 2026, combines experimental biology with a machine-learning model that reads the regulatory code flanking microexons — short pieces of genetic material spliced into messenger RNA. When those flanking sequences are disrupted, the model predicts, microexons are included or excluded incorrectly, potentially altering proteins in the developing brain.
Why microexons matter
Microexons are tiny compared with typical exons, making them easy to miss in bulk sequencing pipelines focused on protein-coding changes. Yet they can shift how neurons wire synapses during development. The research team, including co-lead authors at Memorial Sloan Kettering and the University of Toronto, catalogued new microexons and tested their model on whole-genome data from thousands of autistic individuals and neurotypical siblings.
Collaborators at The Hospital for Sick Children in Toronto applied the predictions to autism-associated variants, finding a subset predicted to alter microexon splicing that concentrates among people previously labelled “idiopathic” — meaning no known genetic cause.
UK science infrastructure
The study underscores Britain’s role in large-scale neurogenetics. King’s and the Crick provided expertise in splicing regulation; SickKids contributed clinical cohorts; computational work bridged institutions on both sides of the Atlantic. Funders include UK Research and Innovation channels that support open access publication, though clinical translation will require separate NHS validation studies.
Researchers caution that risk variants are not deterministic diagnoses. Autism is heterogeneous, and environmental factors interact with genetics. The advance is explanatory: it offers a mechanism that could unify scattered rare variants into a coherent biological story.
Clinical horizon
Genetic counsellors may eventually add microexon-aware interpretation to whole-genome reports offered to families in research settings. Before that, replication in independent cohorts and functional experiments in cell models must confirm that predicted splicing changes occur in human tissue. NHS genomic medicine service leaders will watch whether updated pipelines justify re-analysis of stored genomes from undiagnosed patients.
Ethical and data questions
As with any psychiatric genetics finding, stigma and misuse risks require careful communication. The authors emphasise that the work aims to improve understanding and targeted supports, not to screen populations. Data sharing agreements across countries were essential to reach statistical power; Brexit-era data rules have not stopped this collaboration, but future projects depend on stable cross-border research treaties.
Bottom line
For families who spent years without a genetic explanation, the microexon model offers a fresh place to look in the genome’s dark matter. It is not a bedside test yet — but it is a Nature paper that moves autism genetics from “unknown” toward mechanistic clarity, rooted in London-led science and global cohorts.