How genomics is rewriting autism's story—and why this revolution matters for diagnosis, advocacy, and personalized care
The enigma of autism has fueled a scientific quest spanning decades, billions of dollars, and countless hopes. When Jennifer Singh began her research in 2004, autism was diagnosed in 1 in 150 children. By 2014, that number jumped to 1 in 68—yet no single "autism gene" emerged from this massive scientific effort 1 4 .
This paradox lies at the heart of Multiple Autisms, where Singh reveals how our understanding has evolved from a hunt for singular causes to embracing the spectrum's breathtaking complexity.
Recent breakthroughs now confirm what many lived experiences long suggested: autism isn't one condition with a unified genetic origin. Instead, dozens of interacting genes and distinct biological subtypes paint a picture of profound diversity. This article explores how genomics is rewriting autism's story—and why this revolution matters for diagnosis, advocacy, and the future of personalized care.
The early genetic gold rush was fueled by optimism. Parent-led advocacy in the 1990s and 2000s championed genetics research, donating blood samples and clinical data to build massive genomic databases like the MSSNG project. Scientists initially hoped to isolate a definitive genetic culprit, similar to BRCA1 in breast cancer. But as Singh documents through interviews with scientists and families, this quest proved elusive:
"No single gene has been identified despite a billion-dollar, twenty-year effort—and the more elusive the answer, the greater the search seems to become" 1 4 .
The redefinition of autism as a spectrum (ASD) in the DSM-5 broadened eligibility, capturing diverse symptoms under one umbrella 4 .
| Era | Focus | Prevalence | Key Technology |
|---|---|---|---|
| 1990s-2000s | Single-gene discovery | 1 in 150 | Microarray chips |
| 2010s | Genomic interactions | 1 in 68 | Whole-exome sequencing |
| 2020s | Biologically defined subtypes | 1 in 59 | AI-integrated multi-omics |
A landmark 2025 study led by Olga Troyanskaya at Princeton University finally cracked autism's genetic code—not by finding one answer, but by revealing four distinct biological subtypes.
Researchers analyzed 5,000+ children from the SPARK autism cohort, prioritizing those with "multiplex" families (multiple affected siblings) 6 .
Using AI, they mapped 230+ clinical traits—social behaviors, developmental milestones, co-occurring conditions—onto genetic data.
Whole-genome sequencing identified rare inherited and de novo (new) mutations, while machine learning clustered individuals by biological similarity 6 .
| Subtype | % of Cohort | Core Features | Genetic Drivers |
|---|---|---|---|
| Social/Behavioral | 37% | ADHD/anxiety comorbidities; on-time milestones | Late-acting neural synapse genes |
| Mixed + Developmental Delay | 19% | Motor/speech delays; low co-occurring conditions | Rare inherited variants |
| Moderate Challenges | 34% | Milder core symptoms; few comorbidities | Polygenic risk scores |
| Broadly Affected | 10% | Severe delays, epilepsy, mood dysregulation | High de novo mutations |
Source: Troyanskaya et al. (2025), Nature Genetics 6
ADHD/anxiety comorbidities; on-time milestones
Genetic Drivers: Late-acting neural synapse genes
Motor/speech delays; low co-occurring conditions
Genetic Drivers: Rare inherited variants
Milder core symptoms; few comorbidities
Genetic Drivers: Polygenic risk scores
Severe delays, epilepsy, mood dysregulation
Genetic Drivers: High de novo mutations
Parallel research on the FRRS1L gene illustrates how subtype-specific mechanisms work. This gene, critical for social memory, carries rare variants in families with social interaction challenges. CRISPR-engineered mice with FRRS1L knockouts showed isolated social novelty recognition deficits—mirroring human social challenges without other ASD traits 3 .
Modern autism research relies on integrated tools bridging computation, genomics, and clinical observation.
| Tool | Function | Key Study/Project |
|---|---|---|
| CRISPR-Cas9 | Gene editing in cellular/organoid models | FRRS1L social memory study 3 |
| Perturb-Seq | Large-scale screening of gene function in neurons | Harvard/Broad Institute 7 |
| SPARK Cohort | 43,000+ participants; genomic + trait data | Princeton subtypes study 6 |
| SFARI Gene Database | Curated ASD-risk genes (e.g., 913 scored) | Target validation 9 |
| MSSNG WGS Database | Whole-genome sequences of multiplex families | FRRS1L discovery 3 |
Singh's work sounds a crucial cautionary note: while genomics advances, its real-world impact on autistic individuals remains uneven. Parent advocates often drove research agendas, while autistic adults expressed skepticism about reducing identity to genetics 4 5 .
Genetic testing currently explains only ~20% of cases, leaving many without answers 6 .
Subtype-specific treatments (e.g., FRRS1L-targeted therapies) are emerging, but most drugs only manage symptoms like irritability 9 .
New models prioritize autistic input, ensuring research addresses lived needs 4 .
"Understanding genetic causes could tell families, when their children are young, what symptoms to look for, which treatments to pursue, and how to plan" — Jennifer Foss-Feig, Simons Foundation 6 .
The era of "multiple autisms" is more than a scientific paradigm shift—it's a validation of diversity. As genomics reveals distinct biological narratives, we move closer to personalized supports: early interventions tailored to genetic subtypes, therapies targeting specific pathways like FRRS1L, and diagnostics that map individual trajectories.
Yet Singh's insights endure: autism's meaning extends beyond genomes. In the words of an autistic adult she interviewed, "It's not about fixing us. It's about understanding why we experience the world this way" 4 . As science unravels autism's spectrums, that ethos—centering dignity alongside discovery—may prove its most transformative breakthrough.
"The ability to define biologically meaningful autism subtypes is foundational to precision medicine for neurodevelopmental conditions."
— Natalie Sauerwald, Flatiron Institute 6