Abstract
Introduction: Psoriasis is a chronic immune-mediated inflammatory skin disease with a significant global burden. Current risk assessment lacks integration of proteomic data with genetic and clinical factors. This study aimed to develop a plasma proteomics-based risk score (ProtRS) to improve psoriasis prediction.
Methods: Using data from 53,065 UK Biobank (UKB) participants (1,122 psoriasis cases; 51,943 controls), we integrated 2,923 plasma proteins, polygenic risk score (PRS), and seven clinical risk factors. The Least Absolute Shrinkage and Selection Operator (LASSO) algorithm with 10-fold cross-validation identified stable proteins for ProtRS construction. Population Attributable Fractions (PAFs) for risk factors were calculated.
Results: LASSO regression identified 26 highly stable proteins forming ProtRS-26. ProtRS-26 significantly outperformed PRS and clinical risk factors alone. Combining ProtRS-26 with PRS and clinical factors further improved prediction. Key proteins were enriched in pro-inflammatory pathways and skin-derived. PAF analysis identified hypertension and obesity as major modifiable risk factors.
Discussion: Plasma proteomics significantly enhances psoriasis risk prediction compared to genetic and clinical factors alone. ProtRS-26 provides a robust tool for early screening and personalized prevention.