Autism research has gained a powerful new analytical tool designed to separate the genetic factors children inherit from their parents from the less visible ways parental biology may shape the environments in which children grow up. Developed by University of Virginia School of Medicine geneticist Ziqiao Wang, PhD, and collaborators at Johns Hopkins, the statistical framework could help researchers investigate why genetic risk scores perform differently across populations and how family-level influences contribute to autism and other developmental conditions.
The framework, called PGS-TRI, analyzes “case-parent trios”: a child diagnosed with autism or another condition together with both biological parents. This family-based design gives researchers information that is not available when only the child’s DNA is studied. By comparing the genetic variants a child inherited with the variants present in the parents but not passed down, investigators can estimate whether an association is more likely to reflect a child’s direct genetic susceptibility or an indirect parental influence.
“Traditionally, genetic studies focus on the direct effects of genes passed from parent to child,” Wang said. “However, a child’s health is also shaped by indirect effects—how a parent’s own genetic makeup influences the environment they provide for their child.” A parent’s genetic predisposition to a trait can affect physiology, behavior, health, diet, stress exposure or other aspects of family life, even when the relevant genetic variants are not inherited by the child. These pathways are difficult to measure directly, but family genetic data can provide a statistical route for studying them.
The method builds on polygenic scores, which combine the small effects of many genetic variants across the genome into an estimate of an individual’s inherited predisposition to a disease or trait. Unlike a single-gene test, a polygenic score usually reflects thousands or millions of DNA variants, each contributing a very small amount to overall risk. PGS-TRI extends this approach by modeling the scores of children and both parents simultaneously. This enables researchers to distinguish inherited genetic effects from parental genetic effects that may operate through the prenatal or postnatal environment.
That distinction is especially important in autism research, where risk is influenced by a complex mixture of rare genetic changes, common variants, developmental biology and environmental conditions. No single factor determines whether a child will be autistic, and a polygenic score cannot diagnose autism or predict an individual child’s future with certainty. Instead, such scores are research instruments that can reveal patterns across large populations. By embedding them in a trio-based framework, investigators can ask more precise questions about how risk is transmitted and how family circumstances may interact with inherited biology.
For their initial analysis, the researchers examined more than 18,000 autism case-parent trios drawn from the Simons Foundation Powering Autism Research for Knowledge consortium and the Genes and Environment Autism Research Study. The datasets included families from multiple ancestral backgrounds, allowing the team to test how consistently polygenic scores captured genetic associations across populations. The analysis provided generally reassuring evidence that autism-related polygenic scores reflected direct genetic contributions, while also revealing important differences in their performance among ancestry groups.
Scores were more accurate for families with ancestry from the Americas, Europe and South Asia than for families of African or East Asian ancestry. The researchers attribute much of this gap to the populations historically represented in the large genome-wide association studies used to construct the scores. When discovery datasets contain more individuals from some populations than others, the resulting statistical signals may transfer unevenly across groups. This limitation is not evidence that genetics is inherently more informative for one population; rather, it highlights how unequal representation can reduce the accuracy and usefulness of genomic tools for underrepresented communities.
The findings also generated clues about indirect parental effects. Maternal genetic susceptibility to certain traits, including obesity and some neurocognitive characteristics, was associated with autism risk in children in the study’s analyses. Such results should not be interpreted as proof that a mother’s genes or lifestyle causes autism. Genetic correlations can arise through multiple biological and social pathways, and observational analyses cannot establish that a specific diet, behavior or prenatal exposure is responsible. The value of PGS-TRI is that it offers a framework for testing these possibilities more rigorously in future studies that combine genetic data with detailed clinical and environmental measurements.
The researchers also used the framework to identify the CADM2 gene as a promising target for further investigation. CADM2 has been linked in previous research to behavioral and neurocognitive traits, and the new analysis suggests that it may warrant closer study in relation to autism biology. The finding does not mean that CADM2 is an established cause of autism or that targeting the gene could immediately prevent the condition. Instead, it provides a lead for laboratory and clinical researchers seeking to understand the molecular networks involved in neurodevelopment. Additional studies will be needed to determine whether the association is reproducible, biologically meaningful and therapeutically relevant.
Wang developed PGS-TRI during her postdoctoral training at Johns Hopkins under the supervision of Nilanjan Chatterjee, PhD, whose appointments include the Department of Biostatistics at the Bloomberg School of Public Health and the School of Medicine. The team’s work, published in Nature Genetics, demonstrates how family-based genetic analysis can move beyond a simple comparison of DNA variants and disease status. By combining inherited and non-inherited genetic information, the approach may help clarify the relationship between nature and nurture in autism, developmental disabilities and other complex childhood conditions. The software is publicly available, offering researchers a way to apply the method to new cohorts and improve the representation of diverse populations in genetic studies.
Subject of Research: Genetic and environmental influences on autism risk, polygenic scores, and parent-child genetic effects.
Web References: https://ziqiaow.github.io/PGS.TRI/ ; https://zenodo.org/records/19339894
References: Nature Genetics, DOI: https://doi.org/10.1038/s41588-026-02601-2
Image Credits: Courtesy Wang lab
Keywords: Autism, autism spectrum disorder, genetics, polygenic risk scores, PGS-TRI, parent-child trios, indirect genetic effects, developmental disorders, medical genetics, neurodevelopment, CADM2, personalized medicine, genetic diversity, Nature Genetics
Tags: autismautism risk factorschildhood developmental conditionsenvironmental influences on autismfamily-based genetic researchgene-environment interactionsgenetic inheritance in developmental disordersgenetic risk scoresindirect parental genetic effectsparent-child genetic analysisPGS-TRI frameworkpopulation differences in genetic studies
