Exploring shared genetic architecture in psoriasis and rheumatoid arthritis: a multi-method approach.

Abstract

Psoriasis and rheumatoid arthritis (RA) are immune-mediated diseases that co-occur clinically, but their shared genetic architecture and the direction of shared effects remain debated. We integrated genome-wide association study (GWAS) summary statistics from FinnGen and UK Biobank using linkage disequilibrium score regression (LDSC), high-definition likelihood (HDL), ρ-HESS, MTAG, CPASSOC, colocalization, and MAGMA. The primary cross-population analysis showed a negative global genetic correlation (LDSC unconstrained: rg = - 0.3004, P = 6.70 × 10⁻³; constrained: rg = - 0.2251, P = 3.00 × 10⁻⁴; HDL: rg = - 0.3307, P = 1.14 × 10⁻³). Ancestry-matched sensitivity analyses produced positive but heterogeneous estimates (FinnGen: rg = 0.2422 to 0.6334; UK Biobank: rg = 0.1294 to 0.3777) that were not uniformly significant. Cross-trait analyses identified 95 independent SNPs and 10 additional shared loci. After allele harmonization, all 10 loci showed opposite MTAG effect directions for psoriasis and RA (two-sided exact binomial P = 1.95 × 10⁻³), indicating directional heterogeneity and possible antagonistic pleiotropy. Colocalization provided evidence for a shared causal signal at two TYK2 variants, rs11085727 and rs144309607 (PP.H4 > 0.8; r² = 0.0805 and D' = 1.0), although these results should be interpreted cautiously because they may reflect distinct correlated variants. MAGMA identified immune-related tissue enrichment for psoriasis but not RA. Collectively, these findings indicate that psoriasis and RA share genetic loci, but the global direction of genetic correlation is ancestry- and data-source-dependent and that locus-level effects include opposing directions. TYK2 remains a plausible shared target, but its therapeutic implications require direction-aware fine-mapping and functional validation.