A Residual Approach to Estimate Biological Age from Gompertz Modeling.

Area of Science:

• Gerontology
• Biostatistics
• Computational Biology

Background:

• Biological age (BA) and its residuals quantify aging but often lack robustness and clinical applicability.
• Current methods struggle with heterogeneous populations and direct clinical derivation of age residuals.

Purpose of the Study:

• Introduce the Gompertz law-based residual (GOLD-R) framework for direct estimation of biological age residuals.
• Optimize GOLD-R for cross-sectional data and demonstrate its robustness across multiple datasets and populations.

Main Methods:

• Trained GOLD-R on DNA methylation data (EWAS Data Hub) and evaluated against epigenetic clocks.
• Applied GOLD-R to UK Biobank proteomics data for organismal and organ-specific aging measures.
• Utilized NHANES and HRS clinical biomarker data to compare GOLD-R residuals with epigenetic and phenotypic clocks.

Main Results:

• GOLD-R outperformed established epigenetic clocks in predicting mortality in a pan-cancer dataset.
• GOLD-R aging measures were more robust than conventional approaches in forecasting diseases and mortality using proteomics data.
• GOLD-R residuals from clinical biomarkers surpassed epigenetic and phenotypic clocks in predictive performance.

Conclusions:

• The GOLD-R framework is a robust algorithm for biological age estimation.
• GOLD-R provides a practical tool for both research and clinical applications in aging studies.