Autism is strongly hereditary. Twin studies consistently estimate that genetics account for 64% to 91% of the likelihood of developing autism spectrum disorder, making it one of the most heritable neurodevelopmental conditions. But “hereditary” doesn’t mean a single gene passes from parent to child like eye color. The genetic picture is far more complex, involving hundreds of genes, spontaneous mutations, and interactions with environmental factors.
What Twin Studies Reveal About Heritability
The strongest evidence for autism’s genetic basis comes from twin studies. When one identical twin has autism, the other twin is far more likely to also have it compared to fraternal twins, who share only half their DNA. A meta-analysis combining data from multiple twin studies estimated heritability at 74% under one statistical model, rising to 93% under another. The range depends on how strictly autism is defined and what prevalence rate researchers assume, but every credible estimate lands well above 50%.
Shared environment, meaning factors both twins experience equally like prenatal conditions or household environment, accounts for a smaller slice. Depending on the model, shared environmental effects explain anywhere from 7% to 35% of autism liability. These numbers make clear that while environment plays a role, the genetic contribution is substantially larger.
How Autism Runs in Families
If you already have one child with autism, the chance of a younger sibling also being diagnosed is about 20%, based on data from the Baby Siblings Research Consortium tracking over 1,600 younger siblings. That’s roughly a 1-in-5 chance, compared to roughly 1 in 44 in the general population. An earlier study from the same consortium found an 18.7% recurrence rate, so the risk has remained consistent across research samples.
This familial recurrence doesn’t follow a simple pattern. Some families have multiple children on the spectrum while others have just one. The reason ties back to the complicated genetics underlying autism: parents can carry combinations of risk-associated gene variants without ever being diagnosed themselves, then pass enough of those variants to a child to cross a threshold.
Hundreds of Genes, Not Just One
There is no single “autism gene.” Researchers have identified dozens of genes where rare mutations significantly raise autism risk, including CHD8, SHANK3, SYNGAP1, DYRK1A, PTEN, and ARID1B. Mutations in any one of these can have a strong effect on the individual who carries them, but each accounts for only a tiny fraction of all autism cases.
The majority of genetic risk comes not from these rare, high-impact mutations but from common gene variants scattered across the genome. Each variant individually has a tiny effect, but collectively they add up. This is called polygenic risk, and common variants distributed across many genes account for at least 20% of autism liability at a population level. Research on over 6,400 families showed that parents preferentially pass along these common risk variants to children who develop autism, confirming that everyday genetic variation contributes meaningfully alongside rare mutations.
What makes this even more complex: these influences are additive. A child who carries a strong spontaneous mutation still gets additional risk layered on from common polygenic variation. The genetic pathways involved appear to be at least partially distinct from one another, which helps explain why autism presents so differently from person to person.
Spontaneous Mutations That Aren’t Inherited
Not all genetic causes of autism are passed down from parents. About 3% to 5% of autism cases involve de novo mutations, meaning new genetic changes that appear for the first time in the child and weren’t present in either parent’s DNA. These can be deletions or duplications of DNA segments, or mutations that disrupt how a single gene functions.
De novo mutations are more common in “simplex” families, where only one child has autism and there’s no broader family history. By definition, these spontaneous mutations don’t contribute to the heritability estimate, since they aren’t shared between family members. They’re essentially genetic but not inherited, an important distinction.
One factor that increases the likelihood of de novo mutations is parental age. A study controlling for socioeconomic status and maternal age found that children of fathers aged 40 or older were 5.75 times more likely to have autism than children of fathers under 30. Advancing maternal age did not show an independent association once paternal age was accounted for. Older sperm accumulate more random mutations with each cell division, which likely explains this pattern.
Why Autism Is More Common in Boys
Autism is diagnosed about four times more often in boys than girls, a ratio that has stayed remarkably stable since autism was first described in the 1940s. The leading explanation is called the female protective effect: girls appear to require a greater number or larger magnitude of genetic risk factors before crossing the threshold into a diagnosis.
Supporting this theory, studies find that girls who are diagnosed with autism tend to carry more severe or more numerous genetic mutations than boys with autism. This suggests that female biology provides some buffer against the condition, possibly related to hormonal, chromosomal, or other sex-linked differences. It also means that some girls with significant genetic risk may show subtler traits that go unrecognized, which has fueled growing awareness that autism in girls is likely underdiagnosed.
The Role of Environment
Genetics alone doesn’t account for 100% of autism risk. Environmental factors interact with genetic susceptibility, and researchers increasingly view autism through a gene-environment interaction lens. These environmental influences can be prenatal (infections during pregnancy, exposure to certain medications, complications during birth) or postnatal, and they may alter how genes are expressed without changing the DNA sequence itself, a process called epigenetics.
Animal studies have demonstrated this interaction directly. Mice carrying genetic deletions linked to autism showed fewer behavioral abnormalities when raised in enriched social environments, suggesting that environment can dial the severity of genetic risk up or down. In humans, the interplay is harder to study, but the twin data showing that shared environment explains 7% to 35% of risk confirms that non-genetic factors matter, even if genetics dominates.
Genetic Testing for Autism
Genetic testing can sometimes identify a specific cause of autism, though it won’t find one in every case. The American College of Medical Genetics recommends chromosomal microarray analysis as the first-line genetic test for children with autism. This test detects small deletions or duplications of DNA that standard chromosome tests miss. Fragile X testing is also commonly recommended, since Fragile X syndrome is the most common single-gene cause of autism, identified in about 2% of cases.
For girls with autism who also show developmental regression, testing for mutations in the MECP2 gene (associated with Rett syndrome) is typically suggested. Whole exome sequencing, which reads the coding portions of all genes, is becoming more accessible but isn’t yet standard for every child with autism.
A genetic result can sometimes explain why a child has autism, guide medical monitoring for associated health conditions, and inform family planning decisions. But most children who are tested won’t receive a single clear genetic answer, because their autism stems from the combined effect of many common variants rather than one identifiable mutation.