Simulation of the Metabolic Response to an Interventional Study with New Healthy Beverages by Machine-Learning...

Area of Science:

• Nutritional Science
• Computational Biology
• Pharmacokinetics

Background:

• Interventional trials are resource-intensive and ethically complex.
• Machine learning (ML) offers potential for simulating biological responses.
• Predicting the impact of dietary compounds requires robust analytical methods.

Purpose of the Study:

• To develop and validate ML models for emulating interventional trial outcomes.
• To assess the impact of a maqui-citrus beverage on specific metabolites.
• To evaluate ML model performance in predicting pharmacokinetic changes.

Main Methods:

• Empirical data analysis and preprocessing of beverage consumption data.
• Development of regression models using various ML algorithms.
• Hyperparameter tuning via Bayesian optimization for model refinement.
• Validation of model accuracy using goodness-of-fit (R^2) and error rates (MAE, RMSE).

Main Results:

• ML models achieved high predictive accuracy, with R^2 ≈ 89%.
• Low error rates were observed: MAE ≈ 2% and RMSE ≈ 10%.
• The models successfully predicted changes in flavanone and anthocyanin metabolites in plasma and urine.

Conclusions:

• Machine learning provides a reliable and efficient methodology for emulating interventional studies.
• This approach reduces the need for direct human subject participation.
• The study validates the use of ML in predicting the metabolic effects of dietary interventions.