Abstract
Bipolar Disorder (BD) is influenced by both genetic factors and structural changes in brain regions, along with the impact of external environmental factors on brain network interactions. Conventional genome-wide association studies (GWAS) have identified genetic variants linked to BD but often overlook the spatial interactions and correlations in the brain. In this study, we introduced Spatial Genome-Wide Association Studies with Meta-analysis (Meta-sGWAS), a novel approach to decode the brain areas' interaction patterns and related genetic underpinnings. Based on the genomic and brain MRI datasets from Adolescent Brain Cognitive Development (ABCD) and the UK Biobank (UKB), Meta-sGWAS first revealed significant associations between brain ROIs and BD, as well as the interactions among these ROIs. Notably, the "Right-G_subcallosal" ( P = 9.9 × 10-9 ) and "Left-G_and_S_paracentral" ( P = 3.6 × 10-8 ) were linked to disrupted cognitive and emotional processing in BD. Abnormal connectivity in the right orbitofrontal cortex further highlighted its role in BD-related emotional dysregulation. Meta-sGWAS then incorporated brain ROI interactions to identify genomic traits in a more robust way. From both ABCD and UKB datasets, Meta-sGWAS identified 6 SNPs related to brain-BD interactions. Two SNPs rs1752582 and rs1777305 within the novel identified SVIL gene, linked to brain development, suggest a potential brain-BD connection. Other two SNPs rs12290811 and rs113779084 were found highly enriched in BD-related cells, residing in TENM4 and THSD7A genes, confirming their roles in inhibitory neurons-BD interaction. In summary, Meta-sGWAS integrates brain spatial interactions into genetic association studies, enhancing the accuracy of identifying significant genetic variants and elucidating their interactions. This approach provides valuable insights and potential biomarkers for understanding the brain mechanisms underlying BD pathogenesis.