Abstract
Background
Reduced social connectedness and impaired social functioning are observed across diverse social, psychiatric, and neurodevelopmental phenotypes, but their underlying multivariate genetic architecture remains unclear.Methods
We integrated GWAS summary statistics for Feeling lonely, social deprivation, major depressive disorder, generalized anxiety disorder, and autism spectrum disorder using Genomic SEM to characterize their shared genetic structure. A common-factor GWAS was subsequently performed to identify variants associated with the shared genetic component. Genome-wide significant signals were further characterized using functional annotation, gene-based analysis, statistical fine-mapping, sCCA-TWAS and FOCUS, together with gene-set, tissue and cell-type enrichment and partitioned heritability analyses.Results
The five phenotypes were captured by a common genetic factor with acceptable approximate fit (CFI = 0.963, SRMR = 0.072), with the strongest loadings observed for MDD, Feeling lonely, and GAD. The common-factor GWAS identified 40 genome-wide significant loci, including 25 that reached genome-wide significance only in the multivariate analysis relative to the constituent GWASs. Fine-mapping prioritized six variants with high posterior support across the SORCS3, TENM2, SOX5, and MDGA2 regions, including two SORCS3 variants supported by both SuSiE and FINEMAP. sCCA-TWAS identified 81 Bonferroni-significant expression-feature associations, and FOCUS prioritized 11 genes, seven of which were also significant in sCCA-TWAS. Functional analyses converged on neurodevelopment, synaptic organization, GABAergic and glutamatergic processes, brain tissues, fetal-brain chromatin, and evolutionarily conserved and regulatory genomic regions.Conclusion
These findings identify a shared genetic dimension relevant to social connectedness and functioning across social, psychiatric, and neurodevelopmental phenotypes. Its polygenic architecture converges on neural developmental, synaptic, and regulatory genomic processes.