For a long time, autism has been talked about as if it’s a single condition, just varying in “severity”. But that framing has never really fitted reality.
Some autistic people speak early, read obsessively, and struggle mainly with social nuance. Others don’t speak at all, have intellectual disability, seizures, or profound developmental delays that are obvious in infancy. Calling both of these experiences “the same condition, just on a spectrum” explains very little.
Over the last few years, autism research has been moving in a different direction: autism looks less like one thing on a line, and more like several overlapping pathways that can lead to very different developmental outcomes.
Once you look at it that way, genetics starts to make more sense.
Traits vs disruption: two things hiding under one label
One useful way to think about autism is to separate two overlapping processes that often get bundled together.
1. Autistic traits that run in families
These include things like:
intense focus, sensory sensitivity, preference for structure, literal or detail-oriented thinking, social differences.
Genetically, these traits are polygenic — shaped by hundreds or thousands of common gene variants, each with a very small effect. They tend to run strongly in families and are often associated with strengths as well as difficulties.
This is the form of autism many late-diagnosed adults recognise in themselves and their relatives.
2. Early neurodevelopmental disruption
Alongside that, there is another route into autism that looks very different. Some autistic children show:
Significant language delay, intellectual disability, motor coordination problems, epilepsy, very early developmental divergence.
These presentations are much more often associated with rare, high-impact genetic variants — including de novo mutations (new mutations not present in either parent) or large structural changes in DNA. These variants tend to act very early in brain development and can have broad effects.
Both routes can result in an autism diagnosis, but they are not biologically identical.
Why does genetic concentration matter?
Now here’s where things get interesting.
Autistic traits don’t exist in isolation. People with strong autistic traits are more likely to:
partner with others who share similar traits (assortative mating), come from families where those traits are common, cluster in particular lineages or populations over time.
When that happens, polygenic autistic traits can become highly concentrated within a family line.
On their own, these traits don’t necessarily produce severe disability. In fact, in many environments they may be neutral or even advantageous. But when overall genetic “load” increases — especially if it coincides with rare high-impact variants — the developmental outcome can change qualitatively, not just quantitatively.
Instead of “more traits”, you may see:
Earlier onset, broader developmental impact, higher support needs.
In other words, there may be a threshold effect: below it, autistic traits express as a neurotype; above it, development is disrupted more globally.
How does this fit with recent autism subtype research?
Recent large-scale studies are starting to support this kind of model.
Using detailed behavioural, cognitive, and medical data, researchers have identified distinct autism subtypes rather than a single continuum. These subtypes differ not just in outward presentation, but in:
Patterns of genetic variation, which genes are involved, when in development those genes are most active.
Some subtypes appear more tied to early developmental processes and broader neurological effects, while others align more closely with later-emerging social-communication differences and cognitive styles.
Other work has found meaningful differences between early-diagnosed autism and later-diagnosed autism, again suggesting different biological routes to the same diagnostic label.
All of this points away from the idea that “severe autism is just mild autism turned up louder”.
What this doesn’t mean.
It’s important to be clear about what this way of thinking doesn’t imply.
Not all high-support-needs autism comes from family-level genetic concentration. Many cases involve one-off genetic events. Genetics is not destiny. Environment, prenatal factors, medical complications, and early support all matter enormously. Subtyping should never be used to stigmatise families or police reproduction. Its value lies in understanding, not control.
The goal isn’t to divide people into “good” and “bad” autism. It’s to stop pretending that one explanation fits everyone.
Why does this shift matter?
If autism really is a collection of overlapping pathways rather than a single condition, then:
Research becomes clearer, support can be better matched to actual needs, families get more accurate explanations, late-diagnosed adults stop being confused by models that never fit them.
We’re slowly moving from “one spectrum” to “many autisms”. Genetics — especially how different kinds of genetic risk stack and concentrate — is likely to be a big part of understanding why.
Plymouth Online Therapy 