Integrating multi-omics analysis reveals shared genetic architecture and spatial correlations between heart failure and osteoporosis

Heart failure (HF) and osteoporosis (OP) are common age-related disorders with established epidemiological associations, yet their shared genetic architecture and molecular underpinnings remain poorly defined. We performed an integrative multi-omics analysis to elucidate genetic links between HF and OP. Polygenic overlap was quantified using the bivariate causal mixture model (MiXeR), while shared loci were mapped via conditional/conjunctional false discovery rate (cond/conjFDR) analysis. Transcriptome-wide association studies (TWAS) with MetaXcan, Bayesian colocalization (eCAVIAR, fastenloc), and tissue enrichment analyses were applied to identify candidate genes and regulatory contexts. Spatial mapping of disease-associated cells was conducted using Genetically informed spatial mapping of cells for complex traits (gsMap). MiXeR estimated that approximately 447 ± 114 variants had non-null effects on both traits, corresponding to 52% of the estimated non-null variant count for OP and 21% of that for HF. Despite this extensive polygenic overlap, the genome-wide genetic correlation was weakly negative (rg = -0.091, SE = 0.024, P = 1.07 × 10 − 4 ), and the estimated correlation of effect sizes within the shared non-null component was also negative (rho_beta = -0.295). Cond/conjFDR analysis identified five pleiotropic loci and three novel HF loci. TWAS and colocalization consistently prioritized HOXC6 as a shared transcriptomic candidate across multiple tissues. Spatial transcriptomic mapping using the E16.5 mouse embryonic reference indicated enrichment of HF-associated genetic signals in embryonic muscle regions and OP-associated signals in cartilage and cartilage-primordium regions, with GRB10 emerging as a shared candidate gene. Our study delineates a shared genetic architecture and distinct spatial correlations between HF and OP. HOXC6 and GRB10 were prioritized as candidate shared genes by transcriptomic colocalization and spatial mapping analyses, respectively. These findings generate testable hypotheses regarding potential regulatory links between cardiovascular and skeletal traits and highlight candidate genes and tissues for future functional validation.