Integrating multi-source data and machine learning to Decipher the psoriasis-COPD comorbidity.

The epidemiological and molecular associations between psoriasis and chronic obstructive pulmonary disease (COPD) remain incompletely elucidated. To explore this association and shared mechanisms, this study integrated data of the National Health and Nutrition Examination Survey (NHANES) 2003–2014 (n = 17,416), assessing this association via multivariable logistic regression and subgroup analysis. Transcriptomic data of psoriasis (skin tissue) and COPD (alveolar macrophages) were retrieved from the Gene Expression Omnibus (GEO) database. Candidate biomarkers were identified via differentially expressed gene (DEG) analysis, weighted gene coexpression network analysis (WGCNA), and machine learning [Random Forest (RF) and least absolute shrinkage and selection operator (LASSO)], followed by validation of their diagnostic efficacy. In the fully weighted and adjusted model, no statistically significant association was found between psoriasis and COPD (OR = 1.25, 95% CI: 0.93–1.68, p = 0.14), although trend-level associations were observed among smokers, individuals with hypertension, and those with unstable marital status. We identified 85 shared differentially expressed genes (DEGs), enriched in inflammatory pathways such as the chemokine signaling pathway, and screened three candidate genes (UCK2, P4HA1, and HIBADH). A RF diagnostic model based on these genes achieved Area Under the Curves (AUCs) of 0.935 for psoriasis and 0.962 for COPD in external validation sets. These findings suggest that the comorbidity between psoriasis and COPD may be influenced by risk factors such as smoking and hypertension, as well as shared inflammatory pathways and differentially expressed genes (DEGs) regulation. Psoriasis could serve as a potential window for early COPD screening and provide novel cross-disease therapeutic targets.