Autism spectrum disorder has one of the strongest genetic components of any neurodevelopmental condition, with heritability estimates ranging from roughly 64% to 91% based on twin studies.1PubMed Central. Heritability of autism spectrum disorders: a meta‐analysis of twin studies But the genetics are not simple. No single “autism chromosome” or “autism gene” explains most cases. Instead, the picture involves dozens of chromosomal regions, hundreds of individual genes, and an architecture where both rare mutations and common inherited variation combine in ways researchers are still working to untangle.
How Twin Studies Established the Genetic Basis
The clearest evidence that autism is heavily influenced by genetics comes from comparing identical twins, who share virtually all of their DNA, with fraternal twins, who share about half. A meta-analysis of twin studies found that identical twins matched for an autism diagnosis at a rate of about 0.98, while fraternal twins matched at roughly 0.53 to 0.67, depending on how broadly autism was defined.1PubMed Central. Heritability of autism spectrum disorders: a meta‐analysis of twin studies Those fraternal-twin numbers are higher than you would expect from genetics alone, which hinted early on that shared family environment plays some role too. A California twin study found concordance rates for broader autism of about 77% in identical male twin pairs and 31% in fraternal male pairs, and estimated that shared environment accounted for a meaningful portion of the variance alongside genetics.2Archives of General Psychiatry. Genetic Heritability and Shared Environmental Factors Among Twin Pairs With Autism The overall takeaway, though, is that genetic factors dominate. The question is what kind of genetic variation is involved.
Copy Number Variations and Structural Changes
One of the most productive lines of research has focused on copy number variations, or CNVs. These are stretches of DNA that are either deleted or duplicated compared to a typical genome. Everyone carries some CNVs, but people with autism carry a higher burden of rare ones, particularly deletions. A study using microarray technology found that a larger proportion of autistic individuals carried at least one rare CNV compared to controls (about 90% versus 82%), and that the difference was driven mainly by deletions. When the analysis was narrowed to genes already linked to autism, the excess of deletions in cases became even more striking.3Scientific Reports. An integrated analysis of rare CNV and exome variation in Autism Spectrum Disorder using the Infinium PsychArray
Many of these CNVs arise freshly in the affected person rather than being inherited from a parent. An early landmark study found confirmed new CNVs in about 10% of people with sporadic autism (no affected relative), compared to roughly 1% of controls.4PubMed Central. Strong association of de novo copy number mutations with autism A separate large family study confirmed a major contribution from these new-arising deletions and duplications but also found that inherited ultrarare duplications, passed down from an unaffected parent, can contribute to risk.5PubMed. Rare de novo and transmitted copy-number variation in autistic spectrum disorders That last finding matters because it means a parent can carry a duplication without showing signs of autism and still pass it to a child who does.
A comprehensive microarray analysis of over 400 unrelated autism cases detected unbalanced CNVs in 44% of families. New CNVs turned up in about 7% of families with one affected child and about 2% of families with two or more affected siblings. The study also identified a recurrent CNV at chromosome 16p11.2, appearing at roughly 1% frequency, with features of a genomic disorder.6PubMed Central. Structural variation of chromosomes in autism spectrum disorder
Chromosomal Hotspots Linked to Autism
Certain chromosomal regions turn up repeatedly across studies of autism genetics. Three of the best characterized are 15q11-q13, 22q11.2, and 16p11.2. Each illustrates a different aspect of how chromosomal changes can raise autism risk.
Chromosome 15q11-q13
Duplication of a region on chromosome 15, inherited from the mother, is one of the most common identifiable chromosomal causes of autism. Maternal 15q duplication syndrome (Dup15q) has been linked to roughly half a percent of all autism cases.7PubMed. Phenotypic Description of Autism Spectrum Disorder and Psychopathology in Maternal 15q Duplication Syndrome The clinical picture typically includes intellectual disability, developmental delay, absence of speech, and seizures alongside autism features, though the severity varies widely from person to person.8PubMed Central. Chromosome 15q11-q13 Duplication Syndrome: A Review of the Literature and 14 New Cases This variability in expression makes genetic counseling difficult, particularly in prenatal cases, because the same duplication can lead to vastly different outcomes in different individuals.
The parent-of-origin effect here is important. The 15q11-q13 region is imprinted, meaning certain genes are silenced depending on whether they were inherited from the mother or father. Duplications of the maternal copy tend to cause problems because extra doses of maternally expressed genes in this region disrupt brain development. Paternal duplications of the same region are generally less severe.
Chromosome 22q11.2 Deletion
The 22q11.2 deletion, sometimes called DiGeorge syndrome or velocardiofacial syndrome, is one of the most common chromosomal microdeletions in humans. While it is best known for heart defects and immune problems, it also elevates the risk of autism. One study using rigorous diagnostic tools found that about 18% of individuals with 22q11.2 deletion met strictly defined criteria for autism.9PubMed Central. Examining the Overlap between Autism Spectrum Disorder and 22q11.2 Deletion Syndrome A systematic review of multiple studies found elevated rates of 10% to 40%, though the higher estimates came from studies relying only on parent-report questionnaires, which tend to overestimate.10Journal of Neuropsychiatry. A Systematic Review of Autism Spectrum in Children and Adolescents with 22q11.DS: Prevalence, Clinical Significance and Relationship with Onset of Psychotic Symptoms There is also an open question about whether the social difficulties seen in 22q11.2 deletion are truly autism or a related but distinct set of social impairments driven by anxiety; one study found that most children who scored in the elevated range on autism observation tools also had high anxiety scores.11PubMed Central. Social Impairments in Chromosome 22q11.2 Deletion Syndrome (22q11.2DS): Autism Spectrum Disorder or a Different Endophenotype?
Chromosome 16p11.2 Deletion
Deletions at 16p11.2 are among the most frequent single chromosomal causes of neurodevelopmental difficulties, occurring in roughly 1 in 2,000 people. About 20% to 25% of people with this deletion meet criteria for autism, and around 60% have motor coordination difficulties.12PubMed Central. 16p11.2 deletion syndrome The deletion affects brain volume across all areas and alters white matter properties, which may contribute to the wide range of cognitive and behavioral effects.
Sex Chromosomes and Autism Risk
Autism is diagnosed roughly three to four times more often in males than in females, and sex chromosome variations offer a unique window into why. Males with an extra Y chromosome (XYY) have strikingly elevated rates of autism, with one research cohort finding about 33% meeting diagnostic criteria.13PubMed Central. Autism Spectrum Disorder in Males with Sex Chromosome Aneuploidy: XXY/Klinefelter syndrome, XYY, and XXYY Rates were even higher in XXYY (about 43%), while males with XXY (Klinefelter syndrome) showed a more modest elevation of about 5% to 10%.
A large genome-first study helped tease apart the effects of extra X versus extra Y chromosomes. An extra X chromosome alone did not significantly raise autism risk: people with XXX were not at higher risk than XX individuals, and people with XXY were not at significantly higher risk than XY individuals. But an extra Y chromosome did raise risk. XXY individuals had about 4.6 times the odds of autism compared to XX individuals, and XYY individuals had about 2.4 times the odds compared to XY individuals. Women with Turner syndrome (a single X and no second sex chromosome) also had about six times the odds of autism compared to XX women, roughly matching the rate seen in XY men.14PubMed Central. A genome-first study of sex chromosome aneuploidies provides evidence of Y chromosome dosage effects on autism risk These findings point to a Y-chromosome dosage effect on autism risk and suggest that having two X chromosomes may offer some protection.
The Female Protective Effect
The observation that autism is far more common in males has led to the “female protective effect” hypothesis: the idea that females require a higher burden of genetic risk factors before developing autism. Evidence for this comes from multiple angles. One study found that mothers of autistic children carry significantly more common polygenic risk for autism than fathers of autistic children, consistent with the idea that women can harbor more risk without being diagnosed themselves.15PubMed Central. The female protective effect against autism spectrum disorder Another study showed that autistic females carry a greater burden of harmful new mutations than autistic males, with about 28% more damaging mutations on average in affected girls.16Translational Psychiatry. Genetic evidence of gender difference in autism spectrum disorder supports the female-protective effect
The picture is not entirely settled, however. A critical review pointed out that several key predictions of the standard female-protective-effect model are not well supported by current data, including predictions about heritability differences, the stability of the sex ratio across diagnostic criteria, and clinical features.17Neuron. Can the female protective effect liability threshold model explain sex differences in autism spectrum disorder The sex difference in autism is real, but the mechanism behind it probably involves more than a single threshold model. Hormonal factors, gene-by-sex interactions, and diagnostic biases may all play a role.
Paternal Age and New Mutations
The number of new mutations in a child’s genome increases with the father’s age at conception. Sperm-producing cells divide continuously throughout a man’s life, and each division creates opportunities for copying errors. One study estimated that a father passes on about two additional mutations per year of age, with the mutation count roughly doubling every 16.5 years.18PubMed Central. Rate of de novo mutations and the importance of father’s age to disease risk
For autism specifically, the risk increase tied to paternal age is modest. A study comparing fathers aged 45 to fathers aged 25 estimated that the older father’s child had about a 9% higher risk of autism, an increase that translates to very small absolute numbers given the low baseline rate.19Nature Communications. Paternal-age-related de novo mutations and risk for five disorders This is worth knowing but should not cause alarm: the vast majority of children born to older fathers do not develop autism.
Rare Variants and Common Variants Working Together
For years, the field debated whether autism was primarily caused by rare, high-impact mutations or by the accumulation of many common genetic variants, each contributing a tiny amount of risk. The emerging answer is both. Genome-wide association studies have identified at least five regions of the genome where common variants reach significance for autism risk, and the aggregate of common variation (measured as a polygenic score) explains a meaningful chunk of heritability. At the same time, rare inherited variants and new mutations in specific genes clearly drive risk in some families.
A study of multiplex families (families with more than one autistic member) found that autistic children who carried rare inherited variants also tended to receive a higher-than-expected share of common polygenic risk from their unaffected parents. This combinatorial effect may explain why some parents carry a rare variant without being autistic themselves: they lacked the additional common-variant burden that pushed their child past a threshold.20PubMed Central. The contributions of rare inherited and polygenic risk to ASD in multiplex families Research on whether these rare and common risk factors interact multiplicatively or simply add up has yielded mixed results. A large population-based analysis found no significant interaction between recurrent CNVs and polygenic scores, suggesting the effects are additive rather than synergistic.21PubMed Central. Evaluating the Joint Effects of Recurrent Copy Number Variants and Polygenic Scores on the Risk of Psychiatric Disorders in the iPSYCH2015 Case-Cohort Sample
Epigenetics and Chromatin Remodeling
Not all of the genetic story is written in the DNA sequence itself. Epigenetic mechanisms, which control when and how genes are turned on or off without changing the underlying code, play a growing role in our understanding of autism. Many of the genes identified as autism risk genes encode proteins involved in chromatin remodeling, the process of tightening or loosening the structure that packages DNA.22PubMed Central. Epigenetics, autism spectrum, and neurodevelopmental disorders Genome-wide analyses have found that the chromatin landscape of the genome influences where new mutations tend to land in autism cases and shapes the DNA methylation patterns that govern brain development.23PubMed Central. Autism genes keep turning up chromatin This means that even when the DNA sequence looks normal, problems with how genes are regulated can contribute to autism risk.
Somatic Mosaicism
Most genetic testing looks for variants present in every cell of the body, the kind inherited from a parent or arising at fertilization. But some mutations happen after fertilization, during the early cell divisions of embryonic development. These somatic mutations end up in only a fraction of the body’s cells, creating a patchwork, or mosaic, pattern. If the affected cells happen to include brain tissue, even a mutation absent from a blood sample could influence neurodevelopment.
Estimates suggest that somatic mutations detectable in standard clinical samples contribute to roughly 3% to 5% of simplex autism cases (families with one affected child and no family history). Additional brain-limited somatic mutations, found only in postmortem brain tissue, have also been identified.24PubMed Central. Somatic Mosaicism and Autism Spectrum Disorder Mosaicism matters practically because it can explain why a genetic test on a blood sample comes back negative even when a genetic change is driving the autism. It also complicates recurrence risk counseling, since a mosaic mutation in a parent’s reproductive cells could be passed to more than one child even if the parent’s own blood test looks clean.
What Genetic Testing Actually Finds
Given this complex genetic landscape, families often wonder what a genetic test can realistically reveal. The answer depends on the type of test and on the clinical features of the individual. Chromosomal microarray analysis, which detects CNVs, yields a molecular diagnosis in roughly 9% of children with autism overall, but the rate varies dramatically. In one study, children with autism and additional physical features (such as unusual facial features or organ differences) had a diagnostic yield of about 25%, while children whose autism was the only clinical feature had a yield closer to 4%.25JAMA. Molecular Diagnostic Yield of Chromosomal Microarray Analysis and Whole-Exome Sequencing in Children With Autism Spectrum Disorder Whole-exome sequencing, which reads the protein-coding portions of all genes, added an additional 8% diagnostic rate in the same study. A separate primary care study found that about 24% of microarray results were abnormal in an autism practice, though only a fraction of those were considered clearly clinically significant.26PubMed. Diagnostic yield of chromosomal microarray analysis in an autism primary care practice: which guidelines to implement?
Broader clinical exome sequencing in a large cohort of children with neurodevelopmental conditions found a molecular diagnosis in 27% overall, but only 16% when the primary diagnosis was autism, compared to over 30% for intellectual disability or global developmental delay. Female sex and the presence of minor physical differences were both associated with higher diagnostic yields.27European Journal of Medical Genetics. Diagnostic yield of clinical exome sequencing in 868 children with neurodevelopmental disorders These numbers underscore that most autistic individuals who undergo genetic testing will not receive a single clear genetic explanation for their autism, which reflects the polygenic nature of the condition rather than a failure of the testing.
Fragile X and Single-Gene Conditions
A small but important fraction of autism cases trace to mutations in a single gene. The best known is Fragile X syndrome, caused by an expanded repeat in the FMR1 gene on the X chromosome.28PubMed Central. FMR1 and Autism, an Intriguing Connection Revisited When the repeat grows long enough, it silences the gene, and the resulting loss of its protein product disrupts synaptic function throughout the brain. About a quarter to a third of people with Fragile X meet criteria for autism. Interestingly, research has found that even autistic individuals without Fragile X show unusual patterns of repeat structure in the FMR1 gene, with fewer stabilizing interruptions than typical, hinting that variation in this gene may contribute to autism risk on a broader spectrum.29PubMed. CGG repeat interruptions in the FMR1 gene in patients with infantile autism
Other single-gene causes include mutations in SHANK3, PTEN, CHD8, and SCN2A, among dozens of others. Each individual gene accounts for a very small slice of all autism cases, but collectively they matter. Many of these genes converge on shared biological pathways, particularly those involved in how brain synapses form and communicate. Synaptic cell-adhesion pathways, which govern how neurons physically connect and signal to each other, are disrupted across multiple different genetic forms of autism.30PubMed Central. The emerging role of synaptic cell-adhesion pathways in the pathogenesis of autism spectrum disorders
Down Syndrome and Autism Overlap
Down syndrome, caused by an extra copy of chromosome 21, is not typically thought of as an autism-related condition, but a meaningful minority of people with Down syndrome also meet criteria for autism. A retrospective study of 562 individuals with Down syndrome found that 13% had a co-occurring autism diagnosis. Those with both conditions were more likely to be male and had higher odds of certain medical complications, including infantile spasms and feeding difficulties.31PubMed Central. Co-occurring conditions in children with Down syndrome and autism: a retrospective study Diagnosing autism in someone with Down syndrome can be challenging because some features, like language delay and social differences, overlap between the two conditions. Studies comparing the two groups have found that standard screening tools sometimes fail to clearly distinguish Down syndrome alone from Down syndrome with autism, particularly in measures of social communication and theory of mind.32PubMed Central. Pragmatic competence in people with dual diagnosis: down syndrome and autism spectrum disorder This diagnostic overlap means that autism in people with Down syndrome is probably both underdiagnosed in some individuals and over-attributed in others.
Why Most Genetic Tests Come Back “Negative”
For most autistic people, genetic testing does not return a single identifiable cause. This puzzles many families, especially given the strong heritability figures. The explanation lies in the difference between heritability and identifiability. The roughly 64% to 91% heritability means that genetic differences between people account for most of the variation in who develops autism. But much of that genetic contribution comes from the combined effect of thousands of common variants, each nudging risk by a tiny amount. Current technology can measure this polygenic component as an aggregate score, but it cannot point to one variant and say “this is the cause.” The identifiable single-gene or chromosomal causes explain only a minority of cases, which is why the diagnostic yield of even the most advanced sequencing hovers around 16% to 27% for autism.
This does not mean genetic testing is pointless. When a cause is found, it can guide medical monitoring (some genetic syndromes carry risks for seizures, heart problems, or immune deficits), inform recurrence risk for future pregnancies, and connect families with syndrome-specific support communities and, increasingly, targeted research trials. Induced pluripotent stem cell technology, which allows researchers to grow patient-specific neurons from a skin or blood sample, is opening new avenues for studying how specific genetic variants alter brain cell behavior and for screening potential therapeutic compounds.33PubMed Central. The Use of Induced Pluripotent Stem Cell Technology to Advance Autism Research and Treatment For families weighing whether to pursue testing, the practical question is less “will we get an answer” and more “what would we do with one,” and the answer to that second question is increasingly concrete.