Plasma proteomics enhances genetic risk prediction and biological understanding of venous thromboembolism

Abstract

Venous thromboembolism (VTE) involves complex genetic and molecular interactions not fully captured by current prediction tools. This study integrated a polygenic risk score (PRS) with plasma proteomics data from 44,138 UK Biobank participants to explore the molecular mechanisms underlying VTE. Associations between PRSVTE and 2,911 plasma proteins were analyzed, identifying 265 significant proteins linked to extracellular matrix organization and transmembrane signaling activity. Cox regressions further identified 354 proteins that significantly associated with incident VTE. Mendelian randomization supported causal relationships for 13 proteins, indicating their potential as therapeutic targets. To improve clinical risk prediction, we developed a protein-based risk score (ProteinRS) using LASSO regression. Here we show that the ProteinRS significantly improves VTE discrimination beyond traditional clinical factors and revealed a three-fold gradient in cumulative VTE risk (Hazard Ratio 0.52 vs 1.49). Our findings demonstrate that integrating plasma proteomics with genetic risk scores provides valuable biological insights and improves prediction of VTE.