What we know about genetics & autism

And why it's urgent that the current state of research is better understood

Photo by Eduarda Olechak on Unsplash

Debate over autism’s causes has been nonstop lately. A lot of misinformation is circling, which can happen with even the best of intentions.

This is especially true when it comes to genetics. The science is complex, the language is opaque, and public conversation often lags far behind what researchers are learning.

With so many competing claims in circulation, I wanted to understand for myself what the research actually says.

I’m not a geneticist, and I haven’t read every study. But I’ve now read dozens of them, and I’ve done my best to translate what I’ve learned into plain English.

Two genetic branches of autism

Autism’s genetic pathways can be put into two groups: the first are de novo mutations, and the second are polygenic variants.

Branch one: De novo mutations

Rare, high-impact, mostly spontaneous mutations

In the 2000s, researchers zeroed in on de novo mutations as a cause of autism.¹

De novo means “new,” signifying that these are wholly new mutations. They were not inherited from the person’s parents.

These mutations are associated with intellectual disability such as severe speech and cognitive delays, as well as seizures and motor delay.

This is probably why they were discovered first: they have greater individual impacts than the polygenic factors we’ll discuss next.

The presence of a de novo mutation doesn’t mean a person will have autism. Researchers believe these mutations increase autism risk about 20-fold. In other words, someone with one of these mutations is about 20 times more likely to be diagnosed with autism than someone without.²

Initially, it was thought that these rare genes were entirely de novo (non-hereditary), but there is a wrinkle.³ Researchers have found some rare variants that are similar to the de novo mutations (in that they contribute to developmental delays), but these variants were inherited. This development somewhat complicates the two-branch story I’m tracing here, but these inherited variants are even rarer still. They remain more of a footnote than a challenge to the broader pattern.

Despite their dramatic effects in individual cases, de novo mutations are relatively rare. They don’t account for the majority of autism.

Branch two: Polygenic variants

Subtle, inherited variants that are found in the general population

In the mid-2010s, researchers discovered a polygenic basis for autism. Detecting these subtler genetic factors required large sample sizes and advanced statistical methods, which weren’t as feasible before.

A 2014 study⁴ articulated the shift that came about from this discovery:

Individual risk-associated genes have been identified from rare variation, especially de novo mutations. From this evidence, one might conclude that rare variation dominates the allelic spectrum in autism, yet recent studies show that common variation, individually of small effect, has substantial effect en masse.

These “common variations, individually of small effect” are the polygenic variants – poly denoting multiple.

Polygenic variants are inherited, in contrast with de novo mutations.

Notably, these polygenic variants are “normally distributed” in the population, “which means that some degree of common variant risk” for autism “is present in all of us.”⁵

(Which supports my first foundational principle: “We all (autistic and non-autistic alike) have the same traits, just to different degrees.”)

This kind of genetic risk accumulates like sediment. Some of the relevant genes are present in everyone walking around today. But if they become concentrated in a single individual, a presentation of autism emerges.

When the branches combine

Some people with autism have both bases — de novo and polygenic.

When this happens, there is a cumulative increase in risk.

For instance, one study found that 20% of autistic people with de novo mutations have contributing polygenic variants that, if they were absent, would have meant no presentation of autism.⁶

When we disentangle the branches

Researchers are becoming better at pinpointing the unique contributions of each genetic branch. Here’s how to understand the differences between the pathways:

De novo mutations:

• Are rarer among autistic people and the general population

• The statistically significant mutations are spontaneous (not inherited from one’s parents)

• Tend to have large, disruptive effects on early development

• Are often associated with more visible disabilities or higher day-to-day support needs

Polygenic variants:

• Are common across the general population

• Can contribute to autism when many such variants accumulate

• Are inherited from one’s parents

• Tend to shape cognition in more distributed, often subtler ways

• May bias development toward a different cognitive style, without necessarily resulting in developmental disruption

Categorizing these differences is not meant to imply a hierarchy. Both pathways shape how autism can look and feel.

As one study notes: “These differences strongly suggest that de novo and common polygenic variation may confer risk for [autism] in different ways.”⁷

Photo by Annie Spratt on Unsplash

New discoveries are being made as we speak

For a variety of reasons, what we think we know now will be adjusted as more research is conducted.

The polygenic variants are difficult to catalog

Although much progress has been made in understanding the polygenic basis of autism, researchers haven’t yet aligned on a definitive set of risk-related variants.

For starters, not all contributing variants have been identified yet. Since each variant by itself presents little risk of autism (it’s the cumulative effect that matters), it’s challenging to detect each contributing factor.

It’s also hard to say whether a variant has a big enough impact to count. It’s an exercise in line drawing. How much impact is enough to justify inclusion in the official set?

Researchers are focused on these issues now. For instance, the Psychiatric Genomics Consortium ASD Working Group is actively working on refining and validating the set of variants that are used to score polygenic risk for autism.

The more we learn, the blurrier the polygenic lines become

Many of polygenic variants associated with autism are associated with other polygenic traits or conditions – like ADHD and schizophrenia. For instance, 25% of polygenic risk factors for autism are shared with schizophrenia.⁸

This makes it impossible to isolate a set of gene variants that are purely related to autism. Instead, the picture is one of gradients between neurological profiles and conditions.

Some research also suggests that the traditional definition of autism warrants a second look. A 2024 study found that the three primary domains of autistic traits (which the researchers defined as social, communication, and restricted behavior) have only modest polygenetic correlation.

According to the researchers, this finding “suggests largely independent genetic effects may affect different autistic traits.”⁹

If true, this could suggest the diagnostic definition of autism is somewhat arbitrary, as it groups together traits that don’t deterministically co-occur.

There is still much to learn about de novo genes, too

All this focus on polygenic research is not to say that de novo mutations are fully figured out, either. Every year, multiple studies identify new de novo mutations that contribute to autism.¹⁰

Researchers are also uncovering new mechanisms for how these mutations contribute to autism, like the 2024 discovery of “butterfly effect” genes.¹¹

Why does all this matter?

It’s important to understand the genetic framework of autism for many reasons.

But one stands out right now: science is under threat.

The current U.S. administration is making drastic cuts to medical and scientific research, including autism research. At the same time, the Secretary of Health and Human Services has claimed, without evidence, that a genetic basis for autism is a “myth” and that pursuing this line of research is a “dead end.”

As I hope this post has shown, there is overwhelming scientific evidence that autism has genetic origins — plural. De novo mutations and inherited polygenic risk factors both contribute. And new discoveries are still being made every year, deepening our understanding of the autistic brain.

If we abandon research now, we risk shutting the door on future breakthroughs.

It can be hard to keep up. Genetic research is intricate and often buried in inaccessible language. But the science is there, whether politicians understand it or not. The more we can bring that knowledge into public view, the harder it becomes to erase.

---

Thanks for reading Strange Clarity, where I write about neurodivergence, cognition, and the hidden architectures of thought.

If this post sparked something:
→ Leave a comment: even a simple “I was here” makes a huge impact.
→ Share it with someone thoughtful. Substack makes it easy.

Prefer fewer emails? You can choose which sections to follow by clicking “Manage subscription” at the bottom of any newsletter.

Strange Clarity is reader-powered. If you’d like to support my work, the best way is to subscribe, comment, or share a post that resonated.

You can also find me on Bluesky: @strangeclarity.bsky.com

With curiosity,

Laura

1. See, for example Sebat, J. et al. (2007). “Strong association of de novo copy number mutations with autism.” Science, 316(5823), 445–449. https://doi.org/10.1126/science.1138659

These early-identified genetic markers included CHD8, SCN2A, SHANK3, and NLGN3.

2. Weiner, D. J. et al. (2017). “Polygenic transmission disequilibrium confirms that common and rare variation act additively to create risk for autism spectrum disorders.” Nature Genetics, 49(7), 978–985. https://doi.org/10.1038/ng.3863

3. Satterstrom, F. K., Kosmicki, J. A., Wang, J., Breen, M. S., De Rubeis, S., An, J. Y., Peng, M., et al. (2020). Large-scale exome sequencing study implicates both developmental and functional changes in the neurobiology of autism. Cell, 180(3), 568–584.e23. https://doi.org/10.1016/j.cell.2019.12.036

In the text I say the inherited versions of the genes contributing to developmental delay remain more of a “footnote”; that’s because the study found that their effects are far weaker and their statistical signal barely reaches significance.

4. Gaugler, T., Klei, L., Sanders, S. et al. Most genetic risk for autism resides with common variation. Nat Genet 46, 881–885 (2014). https://doi.org/10.1038/ng.3039

5. Sandin, S. et al. (2017). “The heritability of autism spectrum disorder.” JAMA, 318(12), 1182–1184. https://doi.org/10.1001/jama.2017.12141

6. Gaugler, T. et al. (2014). “Most genetic risk for autism resides with common variation.” Nature Genetics, 46(8), 881–885. https://doi.org/10.1038/ng.3039

7. Weiner, D. J. et al. (2017). “Polygenic transmission disequilibrium confirms that common and rare variation act additively to create risk for autism spectrum disorders.” Nature Genetics, 49(7), 978–985. https://doi.org/10.1038/ng.3863

8. Weiner, D. J. et al. (2017). “Polygenic transmission disequilibrium confirms that common and rare variation act additively to create risk for autism spectrum disorders.” Nature Genetics, 49(7), 978–985. https://doi.org/10.1038/ng.3863

9. de Wit, M.M., Morgan, M.J., Libedinsky, I., et al. (2024). A Systematic Review and Meta-Analysis: Research Using the Autism Polygenic Score. medRxiv. https://doi.org/10.1101/2024.03.08.24303918

10. For instance, this study published in December 2024 identified new de novo mutations. Gogate, A., Kaur, K., Khalil, R., et al. (2024). The genetic landscape of autism spectrum disorder in an ancestrally diverse cohort. npj Genomic Medicine, 9, 62. https://doi.org/10.1038/s41525-024-00444-6

11. George, Jennifer. “The ‘butterfly effect’ explains the genetics of autism.” Wired (Jan. 29, 2025) https://wired.me/science/butterfly-effect-autism/