Psoriasis (PS) is a chronic immune-mediated skin disorder affecting an estimated 2–3% of the global population. PS commonly co-occurs with cardiometabolic conditions, and reported prevalence of metabolic syndrome (MetS) among people with psoriasis ranges widely. MetS is defined by a cluster of metabolic disturbances—dysglycemia, dyslipidemia, central obesity, and hypertension—and its co-occurrence with PS suggests shared pathogenic mechanisms. Traditional epidemiological approaches have limitations for causal inference; therefore, this study uses a multi-level genomic and spatial framework to characterize shared inherited susceptibility and to localize trait-associated signals to specific tissues and cell types using spatial single-cell transcriptomic data.
The study integrated publicly available GWAS summary statistics. Psoriasis was defined using the FinnGen release R12 endpoint L12_PSORIASIS based on ICD-10 codes L40.0–L40.9. MetS was represented by a Lind UK Biobank phenotype (ebi-a-GCST009602) applying Lind’s harmonized NCEP ATP III criteria, requiring at least three of five components: elevated blood pressure or antihypertensive treatment, elevated serum glucose or antidiabetic treatment, elevated triglycerides, elevated waist circumference (WC), and low HDL-C. Component traits analyzed separately included fasting blood glucose (FBG), HDL-C, triglycerides (TG), WC, and hypertension. Summary statistics were harmonized to hg19/GRCh37 and quality-controlled as described in the source GWASs; variants in the extended MHC were handled per the study protocol.
To estimate global genetic sharing the authors applied multiple complementary methods: linkage disequilibrium score regression (LDSC), GNOVA, and high-definition likelihood (HDL). Using multiple estimators allowed assessment of robustness in the magnitude and direction of genome-wide correlations. MiXeR was used to quantify shared and trait-specific polygenic components, addressing polygenic overlap that may not be captured by correlation alone. These tools together establish evidence for genome-wide genetic correlation and for the extent of overlapping polygenic architecture between PS and MetS traits.
Because genome-wide summaries can mask region-specific effects with opposing directions, the study used LAVA to identify local genetic correlations across the genome. LAVA enables characterization of regional heterogeneity, highlighting loci or genomic regions where the correlation between traits differs from global estimates. This step provided a spatial genomic perspective, revealing region-specific signals contributing to PS–MetS sharing.
At the locus level the authors applied conditional FDR (condFDR) to leverage cross-trait enrichment for discovery and conjunctional FDR (conjFDR) to prioritize variants jointly associated with both traits. PLACO, a composite-null pleiotropy test, complemented these approaches by testing for pleiotropy while reducing dependence on single-locus discovery assumptions. Together, these methods identified and cross-validated shared susceptibility loci and candidate pleiotropic genes between PS and MetS-related traits.
To move from genetic loci to cellular context, the study integrated GWAS signals with spatially resolved single-cell transcriptomic atlases using gsMap. This approach evaluates trait-associated enrichment across annotated embryonic tissues and cell types at single-cell resolution, enabling localization of genetic signals to specific tissue compartments and spatial niches. The analysis aimed to identify tissues and cell populations where PS- and MetS-associated genetic variation is disproportionately represented.
Across analytic approaches the authors identified significant genome-wide genetic correlations and demonstrated polygenic sharing between PS, MetS, and several metabolic components. Local analyses revealed region-specific signals and heterogeneity in correlation direction, while MiXeR quantified overlapping polygenic components that can exist even with modest global correlation.
Locus-level cross-trait analyses (condFDR, conjFDR, PLACO) recovered and cross-validated shared susceptibility loci and nominated pleiotropic candidate genes, consistent with immune–inflammatory and metabolic pathways described in prior studies.
Spatial mapping with gsMap showed trait-specific tissue enrichment patterns. For psoriasis, the strongest enrichment was observed in the epidermis (reported pCauchy = 1.0573 × 10⁻⁴), with additional enrichment in adipose tissue (pCauchy = 1.5366 × 10⁻⁴) and liver (pCauchy = 1.0167 × 10⁻³). Across MetS and its component traits (FBG, HDL-C, hypertension, TG), enriched regions were mainly the liver, adipose tissue, and epidermis. Waist circumference (WC) enrichment was predominantly observed in adipose tissue (pCauchy = 1.7823 × 10⁻⁴) and lacked significant liver or epidermal enrichment in these analyses.
By combining genome-wide and local genetic correlation, polygenic overlap quantification, pleiotropic locus discovery, and spatial single-cell mapping, the study links inherited genetic sharing between PS and MetS traits to specific tissue compartments. The enrichment of PS-associated signals in the epidermis is consistent with skin-intrinsic biology, while overlapping enrichment in adipose tissue and liver suggests shared immunometabolic pathways that may underlie comorbidity. The findings create testable hypotheses for functional studies to interrogate how variant effects in these tissues influence both cutaneous inflammation and systemic metabolic processes.
All GWAS data and statistical software used are publicly available from cited resources (IEU, FinnGen, GWAS Catalog) and the authors provided generated results in the main text and supplemental materials. The analytical framework relies on summary statistics and computational integration; as such, the study does not by itself establish causal mechanisms. Details on sample sizes, ancestry composition, and specific locus-level results are provided in the article’s tables and supplemental files. Any functional or clinical inferences require experimental validation.
Integrating GWAS summary statistics with spatially resolved single-cell transcriptomic data characterized shared genetic architecture between psoriasis and metabolic syndrome–related phenotypes and localized trait-associated signals to the epidermis, adipose tissue, and liver. The multi-level framework links genome-wide sharing, regional heterogeneity, polygenic overlap, pleiotropic loci, and spatial cellular context, providing a roadmap for prioritizing functional follow-up studies to investigate comorbidity mechanisms.