Autism, or autism spectrum disorder (ASD), is now regarded as a neurodevelopmental condition with complex, multifactorial causes. Research shows, however, that genetic factors play a major role.
A large-scale study published in JAMA Psychiatry, involving over two million people across five countries, estimated the heritability of autism at around 80 per cent. A more recent Swedish study also arrived at an estimate close to 80 per cent.
It is important to be careful with this figure, as it is often misunderstood. Saying that heritability is 80 per cent does not in any way mean that ‘80 per cent of a person’s autism comes from their genes’. Heritability is a statistical measure applied to a population: it estimates what proportion of the differences observed between individuals can be attributed to genetic differences within that population and in a given environment.
There is no single gene for autism
This is probably the key point.
In some genetic disorders, a specific mutation in a single gene is sometimes sufficient to explain the majority of the condition. This is generally not how autism works.
Hundreds of genetic variations have been linked, to varying degrees, to ASD and other neurodevelopmental conditions. A major study published in *Nature Genetics*, involving more than 63,000 people, identified dozens of genes strongly associated with autism, and many others with more moderate levels of evidence.
Another study involving over 42,000 people with autism showed that genetic risk can stem just as much from rare variants inherited from parents as from spontaneously arising mutations.
In other words, two people with autism may have very different genetic profiles.
In some, a rare variant with a relatively large effect may be found. In many others, it is likely to be the combination of a very large number of common genetic variants, each of which has a tiny effect on its own.
This is known as a polygenic architecture.
And the two phenomena can also combine: common variants, rare variants and de novo mutations can all contribute together to the likelihood of an autistic profile emerging.
So can autism be ‘inherited’ from one’s parents?
Yes, part of the predisposition can be passed on.
However, this does not mean that a father or mother with autism will necessarily have a child with autism. Conversely, a child may be autistic even though neither of their parents is.
Families nevertheless clearly illustrate this genetic component.
An international study published in 2024 followed 1,605 young children who already had an autistic brother or sister. Around 20.2 per cent of them were themselves diagnosed with ASD by the age of three. The rate was even higher when there were several autistic siblings in the family.
This data should not be taken as a rule for every individual: these families constitute a specific population, monitored specifically because a first child was already autistic.
However, it confirms something that geneticists have long observed: the closer two people are genetically, the greater the likelihood that they will share certain characteristics linked to autism.
Studies carried out on twins point in the same direction. A meta-analysis of twin studies produced estimates of heritability ranging from 64 to 91 per cent, depending on the models used.
What about mutations that do not exist in the parents?
That’s the other fascinating part of the story.
Certain changes to DNA are known as ‘de novo’: they arise in the egg, the sperm or very soon after fertilisation and are therefore not present in the same form in the parents’ DNA.
Some of these mutations can have a significant impact on neurological development.
This is one of the reasons why it is possible for a person to have ASD without their father or mother having exactly the same genetic abnormality.
Conversely, certain variants associated with an increased likelihood of the condition may be present in a parent who is not themselves autistic. Genetics is therefore not a switch that simply operates in ‘yes’ or ‘no’ mode.
So, does the environment play no role at all?
That would be taking things far too far.
The US National Institute of Mental Health currently summarises the evidence as follows: the precise causes of autism are not fully understood, and research points to a combined influence of genes and certain environmental factors involved in development.
In this context, ‘environment’ actually has a very broad scientific meaning. It can include biological phenomena occurring during pregnancy or development and does not simply refer to the child’s upbringing or family environment.
Above all, the traditional dichotomy between ‘genetics’ and ‘environment’ is probably too simplistic. Researchers are increasingly studying how these factors interact.
For example, an environmental factor may not have the same consequences depending on an individual’s genetic makeup.
Why do we sometimes find several people with autism in the same family?
This is precisely what high heritability helps us to understand.
Within a family, certain combinations of genetic variants can be passed down from one generation to the next. These can manifest very differently depending on the individual.
One person may be diagnosed with ASD. Another may only display certain characteristics: sensory hypersensitivities, very intense interests, a need for routines, specific social difficulties or an unusual way of processing information, without necessarily meeting the diagnostic criteria for ASD.
Contemporary genetics also shows that biological boundaries are less clear-cut than the categories used in medicine. The same genetic factors can contribute to different neurodevelopmental profiles, and several studies in particular highlight genetic overlaps between autism and ADHD.
Is there a genetic test to determine whether someone is autistic?
Not in the sense of a simple blood test that returns a ‘positive’ or ‘negative’ result.
The diagnosis of autism remains a clinical one: it is based on the person’s development, behaviour, communication and history.
In certain situations, particularly when an ASD is accompanied by an intellectual disability, epilepsy, congenital abnormalities or other medical conditions, genetic testing may, however, identify a variant or syndrome that could explain part of the clinical picture.
But for a large proportion of people with autism, no single ‘culprit gene’ will be discovered.
## How genetics is ultimately changing our understanding of autism
Perhaps, above all, the way we frame the question.
Research no longer really describes ‘autism’ as a single entity with a single cause. Rather, it reveals **a multitude of different biological pathways that can lead to the profiles we now group under the term ‘autism spectrum’**.
A scientific review published in 2025 summarises the field’s evolution as follows: researchers are now simultaneously exploring rare variants, common variants, regulatory regions of the genome and interactions between these different mechanisms.
So yes: autism has a very significant genetic component and can clearly have a familial dimension.
But ‘genetic’ does not mean ‘predetermined’, ‘inevitably inherited’ or ‘caused by a single gene’.
And it is precisely this remarkable biological diversity that is now beginning to explain why the term *spectrum* is probably more accurate than we had imagined.
