Researchers at Johns Hopkins Bloomberg School of Public Health have developed a breakthrough statistical framework that reveals how a child's genes, their parents' biology, and environmental factors like maternal diet and lifestyle collectively shape autism risk—a finding that could transform how scientists understand and predict developmental conditions.
The new PGS-TRI model, described in Nature Genetics, represents a significant leap forward from traditional genetic risk scores, which have long been limited by a critical blind spot: they work well for people of European descent but lose accuracy dramatically in other populations, particularly those of African ancestry. By analyzing data from more than 18,000 case-parent trios—autistic children and their biological parents—drawn from two major research consortiums, Johns Hopkins researchers and their collaborators from Kaiser Permanente Northern California uncovered a far more nuanced picture of how autism risk emerges.
The framework allows clinicians and researchers to move beyond simple DNA predictions. Instead, they can now examine how a child's genetic profile interacts with their parents' genetic predispositions and lived experiences. The research revealed something striking: maternal genetic susceptibility for traits like obesity and certain neurocognitive characteristics may increase a child's autism risk, even independently of the child's own genetic makeup. But the effects don't stop there—known environmental risk factors during pregnancy, such as complications during gestation, multiply the risk further, layering on top of genetic susceptibility rather than simply adding to it.
"This new framework allowed us to gain novel insights into the complex interplay between genes and environment in developmental conditions such as autism," says Nilanjan Chatterjee, the Bloomberg Distinguished Professor of Biostatistics and Genetic Epidemiology who led the research. His team's work suggests that understanding autism requires looking far beyond a single child's DNA sequence—it demands attention to the biological and lived context of their entire family.
The findings also expose an uncomfortable truth in genetic research: the field's reliance on populations of European ancestry has created tools that simply do not work equally well across humanity. The researchers emphasize that developing more accurate genetic risk scores for autism will require sustained, deliberate effort to collect and curate genetic data from diverse populations, especially people of African descent. Without this work, the tools that promise to predict and prevent disease will remain unequally useful.
Looking forward, the Johns Hopkins team is laying groundwork for a new generation of studies. They plan to expand their framework to analyze data from extended family structures—grandparents, siblings, cousins—and to eventually merge findings from family-based studies with massive population-level genetic databases. They're also beginning to develop separate genetic risk scores for children and adults, which could help predict not only autism but a range of health conditions and lifestyle factors across the lifespan.
The researchers acknowledge that while their findings linking maternal genetic traits to childhood autism are compelling, the study design itself could introduce subtle biases that need to be tested through alternative approaches before the results can be considered fully confirmed. But the broader direction is clear: the future of understanding autism lies not in genes alone, nor in environment alone, but in mapping the intricate dance between them across generations of families.