---
title: "Shared genetic architecture and spatial cellular mapping of psoriasis and metabolic syndrome"
id: "plos-one-9-shared-genetic-basis-and-spatial-cellular-atlas-of-psoriasis-and-metabolic"
canonical_url: "https://medichelpline.com/clinical-feed/plos-one-9-shared-genetic-basis-and-spatial-cellular-atlas-of-psoriasis-and-metabolic"
content_type: "clinical_feed_article"
specialty: "General"
source_name: "PLOS ONE (Medicine)"
source_url: "https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0358472"
published_at: "2026-09-15T14:00:00.000Z"
evidence_level: "Journal Feed"
license: "CC-BY-NC-4.0 / Informational Use"
---
# Shared genetic architecture and spatial cellular mapping of psoriasis and metabolic syndrome
## Provenance & Clinical Metadata
- **Canonical URL:** https://medichelpline.com/clinical-feed/plos-one-9-shared-genetic-basis-and-spatial-cellular-atlas-of-psoriasis-and-metabolic
- **Specialty:** [General](https://medichelpline.com/clinical-feed/general.md)
- **Primary Source:** PLOS ONE (Medicine)
- **Source URL:** [Original Journal Publication](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0358472)
- **Published At:** 2026-09-15T14:00:00.000Z
- **Evidence Rating:** Journal Feed
## Executive GIST (TL;DR)
- Psoriasis (PS) and **metabolic syndrome (MetS)** frequently co-occur; this study integrates GWAS summary statistics with spatial single-cell transcriptomics to characterize shared genetic architecture and cellular localization. - Multiple complementary genetic methods were applied: LDSC, GNOVA, HDL for genome-wide correlation; LAVA for local correlations; MiXeR for polygenic overlap; condFDR and conjFDR plus PLACO for pleiotropic locus discovery. - GWAS sources included FinnGen (PS) and Lind/UK Biobank MetS definitions; component traits analyzed were fasting blood glucose (FBG), HDL-C, triglycerides (TG), waist circumference (WC), and hypertension. - Global analyses identified significant genome-wide genetic correlations and polygenic sharing between PS, MetS, and individual metabolic components; local analyses revealed region-specific signals with heterogeneous directions. - MiXeR quantified shared and trait-specific polygenic components, detecting polygenic overlap even when global correlations were modest. - Locus-level cross-trait approaches (condFDR, conjFDR, PLACO) identified and cross-validated shared susceptibility loci and candidate pleiotropic genes. - gsMap integrated GWAS signals with spatial single-cell data to assign trait-associated enrichment to specific tissues and cell populations; **psoriasis** showed strongest enrichment in the **epidermis**, with notable enrichment also in **adipose tissue** and **liver**. - Across MetS and components (FBG, HDL-C, hypertension, TG), enriched regions primarily involved **liver**, **adipose tissue**, and **epidermis**; WC enrichment was predominantly adipose tissue–specific, with no significant liver or epidermal enrichment reported. - The integrated framework provides a chain from genetic association to spatial cellular context, supporting generation of testable hypotheses about PS–MetS comorbidity and priorities for functional and clinical validation. - Data and analytic tools used are publicly available (IEU, FinnGen, GWAS Catalog) and the study supplies results in main text and supplemental data; no new participant-level data were reported.
## Clinical Analysis & Structured Key Points
Shared genetic basis and spatial cellular atlas of psoriasis and metabolic syndrome | PLOS One Browse Subject Areas ? Click through the PLOS taxonomy to find articles in your field. For more information about PLOS Subject Areas, click here . Article Authors Metrics Comments Media Coverage Reader Comments Figures Figures Abstract Background Psoriasis (PS) and metabolic syndrome (MetS) frequently co-occur. Characterizing their shared genetic architecture and spatially enriched cellular populations may clarify the context of their co-occurrence and generate hypotheses for functional validation. Methods We integrated genome-wide association study (GWAS) summary statistics for PS, MetS, and five related components with spatially resolved single-cell transcriptomic data. Global and local genetic correlations were assessed using linkage disequilibrium score regression, genetic covariance analysis, high-definition likelihood, and local analysis of variant association. A bivariate causal mixture model quantified polygenic overlap. Conditional/conjunctional false discovery rate and composite-null pleiotropy analyses identified shared susceptibility loci. Finally, gsMap evaluated trait-associated enrichment across annotated embryonic tissues at single-cell resolution. Results Genetic approaches identified significant genome-wide correlations and polygenic sharing between PS, MetS, and its components. Local and cross-trait analyses identified region-specific signals and cross-validated shared loci. gsMap revealed trait-specific tissue enrichment. PS showed the strongest enrichment in the epidermis (pCauchy = 1.0573 × 10 − ⁴), adipose tissue (pCauchy = 1.5366 × 10 − ⁴), and liver (pCauchy = 1.0167 × 10 − ³). Across MetS, FBG, HDL-C, hypertension, and TG, enriched regions mainly involved the liver, adipose tissue, and epidermis. WC enrichment was predominantly observed in adipose tissue (pCauchy = 1.7823 × 10 − ⁴), with no significant liver or epidermal enrichment. Conclusion Integrating GWAS with single-cell transcriptomic and spatial information characterized shared genetic architecture between PS and MetS-related phenotypes and their spatial enrichment patterns. These findings provide a framework for generating testable hypotheses about comorbidity biology and guiding future functional and clinical validation. Citation: Liu G, Gong F, Yang G, Li Y, Ji Y, Chen X, et al. (2026) Shared genetic basis and spatial cellular atlas of psoriasis and metabolic syndrome. PLoS One 21(9): e0358472. https://doi.org/10.1371/journal.pone.0358472 Editor: Vasudevan Ramachandran, MAIWP International University, MALAYSIA Received: April 14, 2026; Accepted: September 1, 2026; Published: September 15, 2026 Copyright: © 2026 Liu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability: All the GWAS data and statistical software used in this study were publicly available (which can be accessed through the following URLs), and all the generated results in this study were provided in the main text and supplemental data. IEU database: https://gwas.mrcieu.ac.uk/ FinnGen: https://r12.finngen.fi/ GWAS Catalog: https://www.ebi.ac.uk/gwas/home LDSC: https://github.com/bulik/ldsc GNOVA: https://github.com/xtonyjiang/GNOVA HDL: https://github.com/zhenin/HDL LAVA: https://github.com/josefin-werme/LAVA MiXeR: https://github.com/precimed/mixer conjFDR: https://github.com/precimed/pleiofdr PLACO: https://github.com/RayDebashree/PLACO gsMap: https://github.com/JianYang-Lab/gsMap FUMA: https://fuma.ctglab.nl . Funding: The author(s) received no specific funding for this work. Competing interests: The authors have declared that no competing interests exist. 1. Introduction Psoriasis (PS) refers to a chronic, relapsing inflammatory skin disorder driven by immune dysregulation, characterized by sharply demarcated erythematous plaques covered with silvery-white scales [ 1 ]. PS is among the most prevalent immune-mediated conditions worldwide, affecting approximately 2–3% of the global population, with an estimated 125 million individuals currently diagnosed [ 1 ]. The clinical presentation of PS is heterogeneous, with plaque PS (chronic plaque type) representing the predominant subtype, accounting for 80–90% of all cases. It typically manifests as well-circumscribed erythematous plaques coated with silvery-white scales and is frequently accompanied by pruritus [ 2 ]. Lesions most commonly arise on the scalp, extensor surfaces (elbows and knees), lumbosacral region, and nails, and a subset of patients exhibits joint involvement [ 3 ]. Evidence indicates that approximately 30% of individuals with PS develop psoriatic arthritis, resulting in joint pain, stiffness, and impaired functional capacity [ 4 ]. Moreover, PS is frequently associated with multiple comorbidities, among which metabolic syndrome (MetS) requires particular attention, with reported prevalence ranging from 14.3% to 50% [ 5 ]. MetS denotes a constellation of metabolic disturbances, including dysglycemia, hyperlipidemia, central obesity, and hypertension [ 6 ]. The genetic correlation between PS and MetS has garnered increasing interest in recent years [ 5 , 7 ]. Recent genetic studies published in 2026 have further extended this evidence. A cross-trait genome-wide association study (GWAS) identified shared loci and pleiotropic genes connecting PS with MetS and several of its component traits [ 8 ]. The associated genes were enriched in immune–inflammatory, transcriptional, autophagic, and lipid–cholesterol pathways [ 8 ]. Complementary analyses linked psoriasis susceptibility to adverse cardiovascular–kidney–metabolic status and genetically predicted unsaturated fatty-acid profiles [ 9 , 10 ]. Together, these findings support shared immunometabolic susceptibility while highlighting heterogeneity across individual metabolic components. Traditional epidemiological investigations, when examining the mechanisms underlying disease coexistence, are often restricted in their causal interpretability due to confounding, sampling bias, and limited follow-up [ 11 ]. GWAS complement epidemiological evidence by identifying inherited genetic associations at scale; however, they rely on assumptions concerning phenotype definition, population structure, linkage disequilibrium, and dataset comparability and do not by themselves establish causal mechanisms [ 12 ]. Technological advances in functional genomics have enabled integrative analytical strategies to investigate how genetic variants influence cellular processes. Because no single method captures all dimensions of cross-trait genetic sharing, we used a complementary, multi-level analytical framework. LDSC [ 13 ], GNOVA [ 14 ], and HDL [ 15 ] were jointly applied to estimate genome-wide genetic correlations under different statistical frameworks, allowing the direction and magnitude of the global estimates to be assessed for robustness. Because global estimates may obscure region-specific signals with opposing directions, LAVA was used to identify local genetic correlations and characterize regional heterogeneity [ 16 ]. MiXeR complemented correlation-based analyses by quantifying shared and trait-specific polygenic components, including overlap that may persist despite modest genome-wide genetic correlation [ 17 ]. At the locus level, condFDR leverages cross-trait enrichment to improve discovery, whereas conjFDR prioritizes variants jointly associated with both traits [ 18 ]. PLACO provides a complementary pleiotropy test under the composite null hypothesis, reducing dependence on a single locus-discovery assumption [ 19 ]. Finally, gsMap integrates GWAS evidence with spatial omics data to localize trait-associated signals to specific cell types and spatial niches [ 20 ]. Together, these complementary methods connect genome-wide sharing, regional heterogeneity, polygenic overlap, pleiotropic loci, and spatial cellular localization. Guided by this rationale, this study applied the multi-level framework to characterize genome-wide and regional genetic sharing between PS and MetS. We further quantified polygenic overlap, prioritized pleiotropic candidate genes, and mapped trait-associated signals to spatially resolved cellular contexts. This integration provided a coherent evidence chain from genetic association to functional and spatial interpretation, generating testable hypotheses about PS–MetS comorbidity and priorities for future functional validation. 2. Materials and methods 2.1. GWAS Data PS was defined in FinnGen release R12 (endpoint: L12_PSORIASIS) as a binary registry-based endpoint using ICD-10 codes L40.0–L40.9 [ 21 ]. Participants were genotyped using Illumina and Affymetrix arrays and imputed using the Finnish population-specific SISu reference panel. The MetS GWAS (ebi-a-GCST009602) applied Lind’s harmonized NCEP ATP III criteria [ 22 ], requiring at least three of five components: blood pressure ≥130/85 mmHg or antihypertensive treatment; serum glucose ≥6.1 mmol/L or antidiabetic treatment; TG ≥ 1.7 mmol/L; WC > 102 cm in men or >88 cm in women; and HDL-C < 1.0 mmol/L in men or <1.3 mmol/L in women. Although this definition shares the five-component structure of the 2009 IDF/AHA–NHLBI framework [ 23 ], it uses different WC and glucose thresholds. Thus, binary MetS refers specifically to the Lind UK Biobank phenotype. FBG, HDL-C, TG, and WC were analyzed as continuous traits, whereas hypertension was analyzed as a binary trait. Further details on dataset sources, sample sizes, trait definitions, summary-statistics versions and formats, and ancestry are provided in Table 1 in S1 Table . All GWAS summary statistics used in this study originated from previously published datasets that had obtained institutional ethical approval, participant consent, and underwent strict quality control. All source GWASs applied sample- and variant-level quality control and cohort-specific genotype imputation. FinnGen used the SISu reference panel, whereas UK Biobank primarily used the HRC and UK10K/1000 Genomes reference panels. In this study, summary statistics were harmonized to hg19/GRCh37, and single-nucleotide polymorphisms (SNPs) with missing or duplicate rsIDs were removed. No additional minor allele frequency (MAF) or P-value filter was applied during data import. Variants within the extended major histocompatibility complex (MHC) region on chromosome 6 (chr6:25,000,000–34,000,000, hg19/GRCh37) were excluded during preprocessing before all downstream analyses; for LAVA, LD blocks overlapping this interval were removed. This conservative preprocessing step was applied to prevent strong psoriasis-associated HLA signals and complex long-range LD from dominating estimates of the shared genetic architecture. Analyses were restricted to individuals of European ancestry. Exact cohort overlap was unavailable, although overlap among UK Biobank-derived datasets was possible. The complete analytical workflow is shown in Fig 1 . Software versions, repository commits, reference panels, key parameters, genomic exclusion regions, and non-default settings for all analytical tools are summarized in Table 2 in S1 Table . Download: PNG larger image TIFF original image Fig 1. Visual overview of the experimental design. The graphic was constructed with BioRender. LDSC, linkage disequilibrium score regression; GNOVA, genetic covariance analysis; HDL, high-definition likelihood; LAVA, local analysis of variant association; MiXeR, bivariate causal mixture model; condFDR/conjFDR, conditional/conjunctional false discovery rate; gsMap, genetically informed spatial mapping of cells for complex traits. https://doi.org/10.1371/journal.pone.0358472.g001 2.2. Global genetic correlation analyses Three complementary approaches, LDSC ( https://github.com/bulik/ldsc ), GNOVA ( https://github.com/xtonyjiang/GNOVA ), and HDL ( https://github.com/zhenin/HDL ), were used to estimate genome-wide genetic correlations between PS and MetS from GWAS summary statistics. LDSC estimates standardized genetic covariance by relating SNP association statistics to LD scores without requiring individual-level data and is robust to sample overlap [ 13 ]. GNOVA uses weighted genome-wide covariance estimation with explicit correction for sample overlap, improving precision in datasets with overlapping samples or heterogeneous structures [ 14 ]. HDL applies full-likelihood optimization to genome-wide LD matrices with eigenvalue-based dimension reduction, reducing standard errors by approximately 60% relative to LDSC [ 15 ]. Agreement across the three approaches was used to assess the robustness of the genetic-correlation estimates. All analyses used summary statistics after excluding the extended MHC region (chr6:25,000,000–34,000,000, hg19/GRCh37). The reported estimates therefore represent genome-wide genetic correlations outside this predefined interval. 2.3. Local genetic correlation analyses The LAVA ( https://github.com/josefin-werme/LAVA ) method initially partitions the entire genome into multiple independent LD blocks according to LD structural characteristics, ensuring that genetic variants within each block remain highly correlated while maintaining genetic independence across blocks [ 16 ]. On this basis, GWAS summary statistics for PS and MetS, together with LD information matrices derived from European ancestry reference populations, are incorporated to model and estimate the local genetic variance–covariance structure for both phenotypes within each predefined genomic region. Local genetic correlation coefficients are subsequently calculated to quantitatively delineate the extent to which specific genomic regions contribute to disease comorbidity. Through statistical testing, LAVA enables the identification of genomic regions showing significant local genetic correlations, thereby characterizing the regional distribution and heterogeneity of genetic overlap between PS and MetS. In comparison with global genetic correlation analyses, the principal advantage of LAVA lies in its capacity to detect region-specific correlation patterns that may be obscured by genome-wide estimates and to prioritize associated genomic regions for subsequent fine-mapping and functional investigation. 2.4. MiXeR MiXeR ( https://github.com/precimed/mixer ) applies a bivariate causal mixture modeling framework to systematically quantify the extent of polygenic overlap between PS and MetS using GWAS summary statistics [ 17 ]. By fitting the distribution of GWAS effect sizes, this method partitions genome-wide genetic variation into three independent components: PS-specific causal variants, MetS-specific causal variants, and variants shared by both conditions. The quantitative parameters for each component are inferred through maximum likelihood estimation, and the Dice overlap coefficient is subsequently computed to measure genetic similarity (ranging from 0 to 1, indicating complete independence to complete overlap, respectively). This framework facilitates precise quantification of shared and disease-specific genetic architecture at the SNP level, thereby providing quantitative evidence for identifying pleiotropic candidate genes contributing to comorbidity. 2.5. condFDR/conjFDR Analysis The condFDR/conjFDR ( https://github.com/precimed/pleiofdr ) framework was implemented as originally described by Andreassen et al. [ 24 ] and subsequently extended and reviewed by Smeland et al. [ 18 ]. This approach incorporates genetic information from a secondary trait to increase the statistical power of the primary trait, thereby detecting association signals that may be missed in conventional GWAS analyses. The analytical procedure proceeded as follows: GWAS summary statistics from both diseases were initially integrated, followed by quality control and LD correction. Conditional quantile–quantile plots were then generated to evaluate the enrichment of genetic signals for the primary trait (e.g., PS) conditional on the secondary trait (e.g., MetS), where stronger enrichment was reflected by greater leftward deviation of the curve. Based on the enrichment distribution, the condFDR for each SNP was computed using the empirical cumulative distribution function, with correction for genomic inflation applied. Statistical independence was ensured by randomly selecting representative SNPs from LD blocks. The analysis was subsequently repeated with the roles of primary and secondary traits reversed, and the larger of the bidirectionally obtained condFDR values was defined as the conjFDR. Shared loci were identified using a conjFDR threshold of < 0.05. Through its conditional analytical structure, this method increases detection power by integrating information from correlated traits and enables identification of pleiotropic genetic signals without requiring consistency in effect direction [ 18 ]. 2.6. PLACO To systematically identify shared genetic loci between PS and MetS, PLACO ( https://github.com/RayDebashree/PLACO ) [ 19 ] was applied in this study. Unlike conventional approaches, PLACO is anchored in a composite null hypothesis testing framework, in which the null holds when a genetic variant exerts no effect on either trait or influences only one; otherwise, the variant is classified as pleiotropic. In implementation, PLACO extracts effect estimates and standard errors for each SNP from the GWAS summary statistics of both diseases and constructs a joint test statistic based on standardized effect values (Z-scores). This statistic incorporates covariance information between the traits to correct for bias introduced by sample overlap. A central methodological feature of PLACO is its avoidance of any requirement for prespecified effect direction consistency; instead, significant loci are detected at a threshold of P < 5 × 10 − ⁸ by optimizing across all possible combinations of effect directions. This framework enables simultaneous detection of both concordant and discordant pleiotropy, thereby increasing the sensitivity of cross-trait association identification. Moreover, this method relies solely on GWAS summary statistics and maintains robust performance under conditions of sample overlap, effect-size heterogeneity, and varying genetic correlation [ 19 ]. The SNP2GENE functional module of the FUMA [ 25 ] platform was used to conduct detailed gene annotation for genetic variants identified through the condFDR/conjFDR and PLACO analyses. 2.7. Candidate-gene annotation and prioritization Variants meeting the conjFDR (< 0.05) or PLACO ( P < 5 × 10 − ⁸) threshold were mapped to genes using the SNP2GENE module of FUMA [ 25 ]. For focused biological interpretation, candidate genes were prioritized using a hierarchical decision procedure rather than a weighted composite score. Genes identified by both conjFDR and PLACO within the same PS–metabolic trait pair were assigned to the primary priority tier. Within this tier, genes were selected for detailed discussion when at least one peer-reviewed experimental or clinical study linked them to PS and at least one linked them to MetS or the corresponding metabolic component. When multiple genes met these criteria, recurrence across trait pairs and nonredundant representation of metabolic components were used as tie-breakers. This procedure was intended to prioritize representative genes for biological interpretation and was not considered evidence of causality. 2.8. Spatial Transcriptomic Enrichment Analysis To explore the spatial enrichment of trait-associated genetic signals and prioritize potentially relevant cellular populations in PS and MetS, the gsMap ( https://github.com/JianYang-Lab/gsMap ) approach was applied [ 20 ]. gsMap integrates spatial transcriptomics (ST) data wi
## Related Clinical Research

- [Climate change worsens health for people with chronic illness](https://medichelpline.com/clinical-feed/stat-news-1-opinion-climate-change-is-making-people-with-chronic-illness-even-sicker.md)
- [GLP-1 Receptor Agonists in Rheumatic Disease: Mechanisms, Evidence, and Clinical Considerations](https://medichelpline.com/clinical-feed/pubmed-42706128.md) (DOI: 10.3760/cma.j.cn112137-20260312-00692)
- [Hyperuricemia and Gout in People Living with HIV: Systematic Review and Meta-analysis Protocol](https://medichelpline.com/clinical-feed/plos-one-6-urate-dysregulation-hyperuricemia-and-gout-among-people-living-with-hiv-a.md)
- [Testosterone Therapy in Men Often Prescribed Inconsistently with Guidelines, ENDO 2026 Study Shows](https://medichelpline.com/clinical-feed/endocrine-news-0-video-testosterone-therapy-in-men-may-be-overprescribed-inconsistent-with.md)
- [Fatty Acid–Binding Protein 4 Inhibition and Psoriasis: Article Content Not Provided](https://medichelpline.com/clinical-feed/frontiers-in-immunology-16-inhibition-of-fatty-acid-binding-protein-4-alleviates-psoriasis-like-skin.md)

## Navigation
- [← Back to General Feed](https://medichelpline.com/clinical-feed/general.md)
- [← All Clinical Specialties](https://medichelpline.com/clinical-feed.md)
## Medical & Regulatory Disclaimer

> [!CAUTION]
> MedicHelpline content is structured for research, educational, and professional discovery purposes. It does not constitute individual medical advice, clinical diagnosis, or treatment recommendations.
> Always verify dosing, contraindications, and regulatory alerts against official product labeling and primary regulatory sources before clinical decision-making.