The age at which someone gets diagnosed with autism varies wildly. Some children are identified before they can tie their shoes. Others don't get a diagnosis until they're teenagers, adults, or sometimes never at all. The usual explanations for this have focused on things like access to healthcare, parental awareness, or how "visible" someone's symptoms are. All of those matter. But new research suggests there's another player in the game that we've been overlooking: genetics.
A major study published in Nature found that when someone gets diagnosed with autism is partially linked to their genetic makeup. And here's where it gets really interesting: early-diagnosed and late-diagnosed autism appear to involve different genes, different biological pathways, and possibly represent genuinely different conditions at the biological level.
Your Genes Have Opinions About Your Diagnosis Age
Here's a number that should make you sit up: common genetic variants account for about 11% of the variation in diagnosis age. That might not sound like a lot until you realize it's roughly the same contribution as sociodemographic and clinical factors like healthcare access and symptom severity.
In other words, your genes have approximately as much influence on when you get diagnosed as whether your parents could afford an evaluation. That's not what most people would have guessed.
Children who get diagnosed early (before age 6) tend to show social and behavioral challenges during infancy. The signs are there from the start. Kids who get diagnosed later, on the other hand, often have higher rates of co-occurring conditions like ADHD and depression. Their autism might be less obvious early on, or it might be getting masked by other diagnoses that capture clinical attention first.
Different Genes, Different Timing
The really surprising finding is about which genes are involved. When researchers looked at the genetic pathways associated with early versus late diagnosis, they found something striking: the genes involved in each group were active at different points in development.
In people diagnosed later, the impacted genes were mostly active after birth. In people diagnosed early who also had developmental delays, the impacted genes were mostly active prenatally, before the child was even born.
As the researchers put it: "There was little to no overlap in the impacted pathways between the classes."
Different genes. Different developmental timing. Different biological trajectories. This raises a genuine question: are we looking at different forms of autism, or even different conditions that we've been grouping under one umbrella?
The Four Subtypes Model
This new work builds on a Princeton/Simons Foundation study that identified four biologically distinct subtypes of autism:
The first subtype, called Broadly Affected, represents about 10% of cases. These individuals have widespread challenges including social communication difficulties, developmental delays, and mood dysregulation. Everything is affected.
The second subtype involves primarily social and behavioral challenges without many developmental delays. People in this group tend to have the latest average diagnosis age, likely because they develop typically in many domains and their autism is less immediately obvious.
The third subtype is autism with developmental delays, and it's characterized by prenatal gene activity patterns. This group has the earliest average diagnosis age, presumably because developmental delays are noticed quickly by parents and pediatricians.
The fourth subtype, representing about 34% of cases, is similar to the social/behavioral group but with less severe presentations. These individuals might fly under the radar for longer or might not get diagnosed at all.
Why This Actually Matters
Understanding that autism subtypes have different genetic architectures could change how we approach screening, diagnosis, and treatment.
Right now, we have one set of diagnostic criteria for autism and one general approach to screening. But a child showing prenatal-origin developmental delays looks very different from a teenager struggling with social anxiety and ADHD symptoms, even if both ultimately meet criteria for autism. Maybe they need different screening tools. Maybe they need different interventions.
This research also challenges a comfortable simplification that's been convenient but probably wrong: the idea that "autism is autism." The findings confirm what many in the autism community have long argued: autism is not a single condition. It's an umbrella term covering multiple distinct biological trajectories that happen to produce some overlapping features.
Validation for Different Journeys
For families who've waited years for a diagnosis, or for adults discovering their autism in midlife, this research offers something meaningful: validation.
A late diagnosis doesn't mean someone had "mild" autism that was barely there. It might mean they had a biologically different form of autism with different developmental characteristics. The late discovery wasn't just about being overlooked or slipping through the cracks. It reflected real biological differences in how their particular form of autism unfolds over time.
Similarly, a child diagnosed at age 3 isn't necessarily "more autistic" than someone diagnosed at 30. They might just have a form of autism that affects development in more visible ways earlier in life.
Understanding that these are different biological paths, not just different degrees of severity, reframes a lot of individual experiences and potentially points toward more personalized approaches to support.
Reference: Ledford H. (2025). Features of autism can affect age of diagnosis - and so can genes. Nature. doi: 10.1038/d41586-025-03180-8 | PMID: 41034670
Disclaimer: The image accompanying this article is for illustrative purposes only and does not depict actual experimental results, data, or biological mechanisms.