A blood test during pregnancy may reveal epigenetic signals tied to autism

Diagnosing autism spectrum disorder before birth has long seemed out of reach. A study published in June 2026 in Communications Biology by Logan A. Williams and colleagues suggests that prenatal cell-free DNA (cfDNA), fragments of fetal and placental DNA that circulate freely in maternal blood during pregnancy, carries methylation patterns that differ in pregnancies later confirmed as high-likelihood for ASD. Critically, this is preliminary, investigational research, and the findings should not be read as a clinical test ready for use.

cfDNA methylation patterns in ASD pregnancies cluster around synaptic genes

The team performed whole-genome bisulfite sequencing on cfDNA extracted from third-trimester maternal plasma. This technique maps DNA methylation across the entire genome at single-nucleotide resolution, reading chemical tags on DNA that can switch genes on or off without altering the underlying sequence. The cohort was drawn from pregnancies with a high likelihood of ASD, defined by a child's ASD diagnosis at age three, maternal obesity, or both.

When Williams and colleagues compared methylation between ASD and control groups, they identified differentially methylated regions (DMRs): stretches of the genome where the pattern of these chemical tags diverged significantly. Those DMRs were enriched for genes tied to synaptic function, the molecular machinery that neurons use to communicate. They also overlapped with genes already implicated in ASD through prior genetic and epigenetic research. DNA methylation has become a productive marker in non-invasive prenatal testing precisely because it differs between maternal blood cells and placental or fetal tissue, making it detectable in circulating cfDNA.

Why these signals matter for early autism intervention

ASD currently goes undiagnosed, on average, until children are between four and five years old, though many families notice developmental differences earlier. The gap between first parental concern and formal diagnosis costs time that could otherwise be used for intervention. Early autism diagnosis creates measurable benefits, including improved adaptive skills and better access to support services, outcomes that depend heavily on acting during the most neuroplastic window of early childhood.

If prenatal cfDNA methylation signals can eventually be validated in larger, more diverse cohorts, they could move the detection window back by years, to before birth. That is the ambition the Williams study gestures toward, even if it stops well short of delivering a ready-made screening tool.

How maternal obesity intersects with fetal epigenetics

One of the more striking elements of the study is the overlap between the cfDNA ASD DMRs and DMRs associated with maternal obesity (MO). Pregnancies affected by maternal obesity showed epigenetic changes in cfDNA that resembled those seen in ASD pregnancies, suggesting the two risk factors may converge on shared molecular pathways during fetal development.

This is consistent with a large body of prior work. A meta-analysis covering more than 3.68 million mother-child pairs across 42 studies found that children born to overweight or obese mothers face significantly elevated neuropsychiatric risks, including for ASD. A separate study in very preterm infants found that maternal obesity was associated with increased likelihood of screening positive for autism and developmental delay at age two. The Williams findings add a potential biological mechanism: maternal obesity may induce epigenetic dysregulation in the fetal compartment that is detectable in circulating cfDNA during the third trimester.

The researchers also found that the cfDNA ASD DMRs recapitulated methylation patterns seen in two other tissue types: ASD placenta and postmortem cortex from individuals with ASD. That consistency across tissues is notable. It suggests the methylation signal in circulating cfDNA is not an artifact of the blood environment but reflects something happening in the developing brain and placenta simultaneously.

Limitations: small cohort, missing mechanism, and replication needs

The study is investigational, and its constraints are real. The cohort is defined by high-likelihood ASD, which means it was enriched for risk at the outset rather than drawn from a general population. How well these DMR signals would perform as predictors in an unselected prenatal population remains unknown.

The researchers also acknowledge that the mechanistic pathway linking these methylation changes to neurodevelopmental outcomes has not been established. Correlation between cfDNA methylation and later ASD diagnosis does not tell us whether the epigenetic changes are causally upstream of the condition, downstream of some shared cause, or simply co-occurring. Epigenetic findings in ASD brain tissue have repeatedly required replication in larger cohorts, and cfDNA studies face the same challenge, compounded by the fact that cfDNA is a mixture of cell types from different tissues.

Additional prenatal environmental exposures, beyond maternal obesity, could interact with genetics during neurodevelopment in ways this study was not powered to detect.

What comes next in prenatal epigenetic ASD research

The authors frame the study as a proof-of-concept that maternal blood cfDNA can serve as a window into fetal brain epigenetics relevant to ASD. The next steps they point toward include mechanistic studies that ask whether these methylation changes actively alter gene expression in the fetal brain, and larger cohort studies that include a broader range of prenatal exposures.

The finding that maternal obesity and ASD converge on overlapping epigenetic signatures also opens a door to intervention research. If pre-conception or early-pregnancy changes in metabolic health can shift the fetal methylome in a measurable direction, that would add molecular specificity to public health recommendations already in place around gestational weight. Environmental exposures including metabolic factors have been shown to produce epigenetic changes that can propagate across generations, raising the stakes for understanding exactly which prenatal windows matter most.

For now, the Williams study is best read as a demonstration that the biology is there to be found in the blood of a pregnant woman, not yet as a clinical roadmap. The cfDNA methylome is a real and consistent signal. Whether it can be turned into a reliable, population-level prenatal screen for autism risk is a question that will take years and considerably larger studies to answer.

Evidence rating

INVESTIGATIONAL — 61/100

Caveats:

• Sample size is not reported in accessible detail; the cohort appears small and limited to high-likelihood ASD pregnancies, reducing generalizability.
• The study shows epigenetic association, not causation — it cannot establish that cfDNA methylation changes cause ASD or that maternal obesity drives the observed patterns.
• Funding source was not disclosed by the authors, so independent verification of financial interests is not possible.
• Findings have not yet been replicated in an independent cohort.

Single observational cohort study using whole-genome bisulfite sequencing of cell-free DNA from third-trimester maternal plasma. The study identifies methylation associations rather than causal relationships, and the cohort is restricted to high-likelihood ASD pregnancies, limiting generalizability. Findings have not yet been replicated in an independent population.

References

1. Logan A. Williams, Ensieh Habibi, Meghan Miller, et al. (2026). Prenatal cell-free DNA methylome detects association with autism and maternal obesity. Communications Biology. https://doi.org/10.1038/s42003-026-10500-9
2. Fabio Coppedè, Utsa Bhaduri, Andrea Stoccoro, Vanessa Nicolì, Eleonora Di Venere, Giuseppe Merla (2023). DNA Methylation in the Fields of Prenatal Diagnosis and Early Detection of Cancers. International Journal of Molecular Sciences. https://doi.org/10.3390/ijms241411715
3. Chiugo Okoye, Chidi M Obialo-Ibeawuchi, Omobolanle A Obajeun, et al. (2023). Early Diagnosis of Autism Spectrum Disorder: A Review and Analysis of the Risks and Benefits. Cureus. https://doi.org/10.7759/cureus.43226
4. Maternal Obesity and Neurodevelopmental Outcomes in Offspring: A Narrative Review – PMC. NIH. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9351052/
5. L C Reynolds, T E Inder, J J Neil, R G Pineda, C E Rogers (2014). Maternal obesity and increased risk for autism and developmental delay among very preterm infants. Journal of Perinatology. https://doi.org/10.1038/jp.2014.80
6. Starnawska Anna, Janecka Magdalena (2023). Editorial: Epigenomic contributions to autism spectrum disorders. Frontiers in Neuroscience. https://doi.org/10.3389/fnins.2023.1177378
7. Epigenetics and Autism Spectrum Disorder: Is There a Correlation? – PMC. NIH. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8900694/