Mapping the Regulatory Effects of Non-Coding Variants in Autism

Research at Stanford University has led to significant breakthroughs in understanding the relationship between non-coding variants and developmental diseases, including autism. Whole genome sequencing has revealed the non-coding genome as a substantial contributor to disease, while attempting to prioritize causal common and rare non-coding variants remains a substantial challenge. By utilizing deep learning models trained on single-cell ATAC-seq across 132 cellular contexts in adult and fetal brain and heart, researchers have produced nearly two billion context-specific predictions. These predictions have been used to distinguish candidate causal variants underlying human traits and diseases, including autism.

Key Takeaways:

• Whole genome sequencing has identified over a billion non-coding variants in humans, with GWAS revealing the non-coding genome as a significant contributor to disease.
• The research team used deep learning models to predict the effects of 15 million variants across 132 cellular contexts in adult and fetal brain and heart, producing nearly two billion context-specific predictions.
• The predictions were used to prioritize candidate causal variants underlying human traits and diseases, including autism, and to identify mutation outliers near syndromic autism-associated genes.
• FLARE, a context-specific functional genomic model of constraint, was developed to prioritize mutations with extreme regulatory effects, which outperformed other methods in prioritizing case mutations from autism-affected families.
• The research demonstrated the potential of integrating single-cell maps with population genetics and deep learning-based variant effect prediction to elucidate mechanisms of development and disease.
• The findings suggest that genetic contributions to neurodevelopmental disorders are predominantly rare, and that prioritizing causal common and rare non-coding variants remains a significant challenge.

Statistics:

• Over a billion non-coding variants have been identified in humans through whole genome sequencing.
• 15 million variants were analyzed using deep learning models trained on single-cell ATAC-seq across 132 cellular contexts.
• Nearly two billion context-specific predictions were produced using the deep learning models.
• FLARE, the context-specific functional genomic model of constraint, was developed to prioritize mutations with extreme regulatory effects.
• The research team identified mutation outliers near syndromic autism-associated genes in autism-affected families.

Sources:

• Mapping the regulatory effects of common and rare non-coding variants across cellular and developmental contexts in the brain and heart, bioRxiv, 2025.
• NewsRx, Findings on Autism Reported by Researchers at Stanford University (Mapping the regulatory effects of common and rare non-coding variants across cellular and developmental contexts in the brain and heart), Mental Health Weekly Digest, November 3, 2025; p 192.