Genes themselves, not the mother's illness, were doing the work
A Danish research team has found a way to distinguish between two fundamentally different pathways through which a mother's health may shape her child's autism risk — whether through the genes she passes down or through the prenatal environment her body creates. Using national registry records and the natural genetic experiment offered by cousin relationships, the researchers sidestepped the need for expensive DNA analysis entirely. The study invites us to consider how much wisdom about biological inheritance may already be encoded in the ordinary facts of family and medical history, waiting for the right question to unlock it.
- Scientists have long struggled to separate whether a mother's illness raises autism risk by passing on genes or by altering the womb environment — a distinction with real consequences for how autism is understood and potentially prevented.
- The Danish team turned to 18,374 autistic children and their extended maternal families, using cousin comparisons as a natural genetic lens to cut through a problem that normally demands costly genomic tools.
- The results split cleanly: conditions like epilepsy pointed to shared genetic inheritance, while false labor and anxiety disorders showed patterns suggesting the mother's genes were reshaping the prenatal environment itself.
- Researchers outside the study are already calling for the method to be applied to other large populations and combined with actual genetic data, signaling that this elegant design may be just the beginning of a broader investigative program.
For years, scientists studying autism have faced a stubborn puzzle: when a mother's health condition correlates with autism in her child, is that because of genes she inherited, or because her condition altered the prenatal environment in ways that shaped the developing brain? The distinction matters enormously — it determines where biological explanations and potential interventions should be aimed. A Danish team led by Magdalena Janecka at NYU Langone Health found a way to answer this question using nothing more than family records and a clever logical design.
The researchers drew on medical registries covering 18,374 Danish children diagnosed with autism between 1998 and 2015, examining 236 different maternal diagnoses recorded in the four years before pregnancy. Their key tool was the cousin. If a maternal condition influences autism through genes, then all cousins on the mother's side should show similar patterns regardless of whether their mothers are sisters or brothers. But if the condition works by changing the prenatal environment, then cousins whose mothers are sisters — who share more genetics and would therefore create more similar wombs — should cluster more tightly together.
The data revealed a clear split. Epilepsy, postpartum hemorrhage, and personality disorders showed equal associations across all cousin types, pointing to direct genetic inheritance. But false labor, recurrent depression, anxiety disorders, and connective tissue disease showed stronger links among parallel cousins, suggesting the mother's genes were shaping her prenatal environment in ways that fed into her child's autism risk.
What struck outside observers was the method's economy. No DNA samples, no sequencing — only the relationships and diagnoses a national registry already holds. Experts including Michael Eriksen Benros, Lisa Croen, and Brian Lee praised the design and pointed toward natural extensions: applying it to other populations, and eventually layering in actual genetic data from cohorts like iPSYCH. The study, posted as a preprint in April, is a reminder that sophisticated questions about genes and environment sometimes yield to surprisingly simple tools.
A team of researchers in Denmark has found a way to answer a question that has long puzzled scientists studying autism: when a mother's health condition is linked to autism in her child, is that connection because of genes she passes down, or because her condition changes the prenatal environment in ways that affect the developing brain?
The answer matters. It shapes how we understand what causes autism and where prevention efforts should focus. But teasing apart these two pathways—direct genetic inheritance versus indirect environmental influence—is notoriously difficult. You need genetic data, large populations, detailed medical histories, and a clever research design. The Danish team, led by Magdalena Janecka at NYU Langone Health, found a way to do it using something simpler: cousins.
The researchers examined medical records for 18,374 children diagnosed with autism in Denmark between 1998 and 2015, along with records for their mothers and all their maternal aunts, uncles, and cousins. They looked at 236 different diagnoses their mothers had received in the four years before pregnancy—everything from epilepsy and high blood pressure to depression and false labor. Then they asked a deceptively simple question: were the autistic children's cousins also more likely to have autism?
The logic is elegant. If a maternal condition affects autism through genes the mother passes to her child, then cousins on the mother's side should show the same pattern regardless of which aunt they're descended from. But if the condition works by changing the prenatal environment—say, by altering hormone levels or nutrient availability in the womb—then cousins with mothers who are sisters (parallel cousins) should be more similar to each other than cousins with mothers who are brothers (cross cousins). Sisters share more genetics than a sister and brother do, so they would create more similar prenatal environments.
What the team found was a split. Some conditions, like epilepsy, postpartum hemorrhage, and personality disorders, showed equal associations with autism across all cousin types. This pointed to direct genetic effects—the genes themselves, not the mother's illness, were doing the work. But other conditions told a different story. False labor, recurrent depression, anxiety disorders, and connective tissue disease showed stronger links to autism in parallel cousins than cross cousins. This pattern suggested indirect genetic effects: the mother's genes were shaping her prenatal environment in ways that influenced her child's autism risk.
The elegance of this approach is that it requires no genetic sequencing, no DNA samples, nothing beyond what a national health registry already contains. Janecka and her colleagues used only family relationships and medical diagnoses. Yet the method can distinguish between two fundamentally different biological pathways—a distinction that would normally require expensive, time-intensive genetic analysis.
Experts outside the study see its potential. Michael Eriksen Benros, a professor at the University of Copenhagen, called it impressive that a registry-based family design could separate direct and indirect genetic effects without using genetic data at all. Lisa Croen, who directs the Autism Research Program at Kaiser Permanente, suggested the strategy should be applied to other large populations with detailed medical records. And Brian Lee, an epidemiologist at Drexel University, noted that the obvious next step would be to layer actual genetic data onto what the family design reveals—something the Danish team is positioned to do, given access to the well-characterized iPSYCH cohort.
The study, posted as a preprint in April, opens a door. It shows that you don't need the most sophisticated tools to answer sophisticated questions about how genes and environment interact. Sometimes the answer is hiding in the relationships between people you already know about.
Citações Notáveis
Using family designs, we can still probe questions about the role of genetics in certain epidemiological associations without genetic data.— Magdalena Janecka, associate professor at NYU Langone Health
This registry-based family design can separate how direct and indirect genetic effects contribute to observed links between maternal health and offspring autism.— Michael Eriksen Benros, professor at the University of Copenhagen
A Conversa do Hearth Outra perspectiva sobre a história
Why does it matter whether a maternal condition affects autism through genes versus through the prenatal environment?
Because the two pathways point to completely different interventions. If it's genetic, you're looking at inheritance patterns and maybe someday genetic screening. If it's environmental, you might be able to modify what happens during pregnancy—manage the mother's condition differently, adjust nutrition, change medications. The mechanism changes everything about what you can actually do.
But couldn't both be true at the same time?
Absolutely. And that's what makes this study clever. It's not saying one or the other. It's saying: here are conditions where we see evidence of direct genetic effects, and here are conditions where we see evidence of indirect effects. Some might have both. The point is to start sorting them out.
How does comparing cousins actually tell you about genetics without looking at genes?
Family relationships are a kind of natural experiment. Parallel cousins—children of two sisters—share the same maternal genes through their mothers. Cross cousins—children of a sister and brother—don't share those maternal genes the same way. If a pattern shows up equally in both, genes probably aren't the main story. If it shows up more in parallel cousins, maternal genetics are likely involved.
What surprised you most in the findings?
That some conditions split so clearly. Epilepsy showed up equally across all cousin types, suggesting it's purely genetic. But depression and anxiety showed the parallel-cousin pattern, hinting that the mother's genetics shape her prenatal environment in ways that matter for autism. It's not one simple story.
Can this method work in other countries?
That's the hope. You need a national health registry with detailed medical records and family relationships tracked over time. Denmark has that. So do several other countries. The method is elegant partly because it doesn't require expensive genetic sequencing—just good record-keeping and the right research question.