Multivariate genetic analysis reveals three distinct pathological dimensions in musculoskeletal disorders

The substantial overlap in symptoms and pathology across musculoskeletal disorders underscores the need to investigate their shared genetic mechanism. Using multivariate genetic modeling, we identify three distinct pathological factors, including a degenerative musculoskeletal disorder factor, a bone mass factor, and an autoimmune disorder factor, which collectively explain 48.27% of the total genetic variance in musculoskeletal disorders. Genome-wide association analyses of these factors identify 795 genomic risk loci, 136 of which are novel, along with 273 candidate genes, 20 of which interact with existing drugs. All three factors exhibit significant heritability and polygenic architecture, while showing low genetic correlations and little overlap in associated loci. Polygenic enrichment analyses demonstrate the potential roles of brain-joint neural axis underlying the degenerative musculoskeletal disorder factor, which is highly enriched in mouse embryonic brain tissue (P = 1.12 × 10^-12) and neuronal cells annotated from a single-cell atlas of human embryonic skeletal development (P = 1.88 × 10^-20), as well as the gut-joint immune axis underlying the autoimmune disorder factor, which is highly enriched in mesenteric lymph node (P = 1.99 × 10^-8) and colonic epithelium (P = 4.36 × 10^-7). The distinct pathological dimensions underlying three factors highlight the importance of genetic-guided stratification for optimizing the co-prevention and co-management of musculoskeletal disorders.