Whole-genome sequencing reveals variant associations with brain imaging phenotypes

Abstract

Whole-genome sequencing (WGS) enables comprehensive discovery of genetic variation underlying brain imaging traits beyond the limits of SNP arrays, particularly for rare coding variants with potentially large effects. Here, we integrated WGS data from 45,755 UK Biobank participants with 4,016 derived imaging phenotypes (IDPs) spanning macrostructure (volumes, cortical thickness/area), microstructure (diffusion metrics), functional connectivity, and perfusion MRI measures. We performed variant-level association testing across 2,751,139 coding SNVs under dominant, genotypic, and recessive models. Then, we conducted gene-based rare-variant collapsing analyses across 18,762 genes using multiple qualifying-variant models to improve power for ultra-rare alleles and enhance biological interpretability. We detected 66,893 significant genotype-IDP associations at the variant level and 184 significant gene-IDP associations from collapsing analyses, including many signals driven by protein-truncating and damaging missense variants with larger effect sizes than common variants. Burden heritability analyses supported a contribution of rare coding variation to population variability in IDPs. This study provided a WGS framework for mapping rare coding variant effects on brain structure and function, advancing mechanistic insight into imaging endophenotypes relevant to neuropsychiatric and neurodegenerative disease.