Multilayer Genomic Characterization of a Shared Genetic Factor Linking Depression-Related Liability and Reduced Physical Function

Background: Depression-related liability is frequently accompanied by reduced physical function, yet the shared genetic architecture linking mood-related traits and physical-function decline remains incompletely characterized. Methods: We applied genomic structural equation modeling to European-ancestry GWAS summary statistics for five constituent phenotypes: depressive symptoms, depression diagnosis, grip strength, appendicular lean mass, and walking pace. A Depression–Physical Function shared genetic factor was constructed as a cross-trait genetic covariance dimension and evaluated using LDSC-based validation and leave-one-trait-out sensitivity analyses. We then performed factor GWAS, FUMA locus annotation, Bayesian fine-mapping, MAGMA gene-based analysis, transcriptome-wide association analysis, pathway enrichment, CELLECT/MAGMA cell-type specificity analysis, partitioned heritability analysis, and gsMap spatial transcriptomic mapping. Results: The shared factor showed good model fit and retained 755,397 quality-controlled variants for downstream analysis. The factor was positively genetically correlated with depression-related traits and negatively correlated with physical-function-related traits. FUMA identified 245 genome-wide significant SNPs, 44 lead SNPs, and 38 genomic risk loci, with 127 positional mapped genes. Fine-mapping prioritized one high-confidence locus. MAGMA identified 19 Bonferroni-significant genes and 326 FDR-significant genes, while TWAS identified 322 FDR-significant expression-associated genes. Integrating FUMA positional mapping, MAGMA gene-level association and TWAS expression-level association prioritized eight convergent genes: TMEM106B, CENPW, DRD2, LRFN5, NCAPG, DCAF16, SGIP1, and FAM120A. Functional enrichment highlighted postsynaptic structure, neuron spine, synaptic plasticity, and synapse organization. CELLECT/MAGMA prioritized brain non-myeloid neurons and glial populations, with additional endocrine-metabolic and immune-hematopoietic signals. Spatial transcriptomic mapping localized top signals to brain and spinal cord regions in the embryonic neuro-muscle reference. Partitioned heritability analysis showed enrichment in conserved, intronic, promoter, and chromatin-related genomic annotations. Conclusions: These findings support a shared polygenic covariance dimension linking depression-related liability with reduced physical-function-related genetic propensity. Downstream analyses prioritized candidate loci, genes, and biological contexts, with enrichment patterns consistent with neuronal, synaptic, and regulatory genomic processes.