Microbiome-Based Aging Clock in a Southeast Asian Population: Development and Validation with Han Chinese Emphasis

Ng, S. M., Tan, S. C., Watcharapichat, P., Lim, J., & Lee, J. W. J. (2025). Microbiome-based aging clock in a Southeast Asian population: Development and validation with Han Chinese emphasis [Poster presentation]. 10th Microbiome R&D and Business Collaboration Congress.

Ever wondered what you gut microbes say about your biological age?

We developed a microbiome-based aging clock tailored but not limited to predict biological age in the Asian population, thus providing valuable insights into the relationship between aging and health.

• Aging clock: A predictive model that estimates biological age using age-related biomarkers.
• Biological age: The years that reflect an individual’s physical condition and how well their body functions.
• Chronological age: The number of years an individual has lived since birth.

METHOD

1. Healthy cohort: No self-reported medical conditions, non-smoking, normal Asian BMI, no antibiotic usage for past 3 months
2. Model training: Reduced features via Recursive Feture Elimination, trained model using different algorithms with 10-fold cross-validation repeated 3 times and eliminated systemic biasfree prediction via an adapted linear regression
3. Model testing: Evaluated model performance by calculating the MAE and R2 across various cohorts
4. Visualisation: Visualised accelerated/ delayed aging across healthy and diseased cohorts

RESULTS

• Our random forest model offers a MAE ranging from 1.79 - 2.78 and R2 of 0.96 - 0.98 when evaluated against different cohorts.
• Healthy individuals exhibited delayed aging, whereas the diseased cohort with metabolic syndrome showed accelerated aging.

CONCLUSION

To the best of our knowledge, our model represents the first microbiome-based aging clock developed in Southeast Asia, and offers the highest accuracy to date in the literature. Further studies are warranted to associate the biological age with frailty and mortality risk.