Prediction of Lung Function in Adolescence Using Epigenetic Aging: A Machine Learning Approach

Area of Science:

• Epigenetics
• Pulmonology
• Computational Biology

Background:

• Epigenetic aging is linked to various health outcomes and diseases.
• Previous research on epigenetic aging and lung function primarily focused on older populations.
• The impact of epigenetic age acceleration on lung function in adolescents remains underexplored.

Purpose of the Study:

• To investigate the predictive power of epigenetic age acceleration (AA) for adolescent lung function.
• To apply machine learning techniques to analyze DNA methylation data and lung function metrics.
• To assess lung health trajectories from childhood to adolescence using epigenetic markers.

Main Methods:

• DNA methylation-based epigenetic age accelerations (AAs) were calculated from 326 samples at ages 10 and 18 from the Isle of Wight Birth Cohort.
• Five machine learning regression models (linear, lasso, ridge, elastic net, Bayesian ridge) were employed.
• Feature selection based on mutual information identified key predictors for modeling FEV1 and FVC at age 18, incorporating AA changes between time points.

Main Results:

• Ridge regression achieved R² = 75.21% ± 7.42% for predicting FEV1.
• Elastic net regression yielded R² = 75.38% ± 6.98% for predicting FVC.
• Both models demonstrated significant predictive accuracy for lung function parameters in adolescence based on epigenetic aging markers.

Conclusions:

• Machine learning, combined with longitudinal tracking of epigenetic age changes, offers a promising approach for assessing adolescent lung health.
• Epigenetic age acceleration is a significant predictor of lung function development during adolescence.
• This study highlights the potential of epigenetic markers for early identification of lung health issues.