ProtPhenoAge: Integrating plasma proteomics to predict Aging-Related disease Risks.

Area of Science:

• Biomarkers and aging research
• Proteomics and computational biology
• Genomics and personalized medicine

Background:

• Plasma proteins offer insights into aging influenced by internal and external factors.
• Proteomics-based aging clocks represent a promising method for quantifying the aging process.

Purpose of the Study:

• To develop and validate a novel proteomics-based aging clock by integrating plasma proteomics with composite biomarkers.
• To assess the performance of the new clock in predicting biological age and all-cause mortality.

Main Methods:

• Utilized UK Biobank data (37,433 participants) with Olink Explore proteomics data.
• Employed Boruta-SHAP to identify PhenoAge-related proteins and trained six machine learning models to develop ProtPhenoAge.
• Conducted Phenome-wide association study (PheWAS), Mendelian randomization (MR), and genome-wide association study (GWAS) to explore associations and identify aging-associated loci.

Main Results:

• The ProtPhenoAge model (XGBoost) strongly correlated with PhenoAge (r=0.96, R²=0.92) and predicted all-cause mortality (AUC=0.76), outperforming existing clocks.
• ProtPhenoAge Acceleration (ProtPhenoAgeAccel) associated with 313 disease phenotypes, revealing more age-independent aging features than previous clocks.
• Identified 10 aging-associated loci via GWAS, including variants linked to epigenetic aging and the APOE gene.

Conclusions:

• ProtPhenoAge quantifies the aging process more effectively than previous clocks by detecting time-independent aging features.
• Genomic and phenomic evidence supports ProtPhenoAge as a reliable tool for aging assessment and research.
• The findings highlight the potential of ProtPhenoAge in advancing aging research and clinical applications.