Predicting the health benefits of Beta-glucan using machine learning
Beta-glucan, a naturally occurring polysaccharide from sources such as oats, barley, yeast, and mushrooms, has gained significant attention for its health benefits, particularly in immune modulation and cholesterol reduction. However, a comprehensive understanding of the relationship between beta-gl...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-09-01
|
| Series: | Franklin Open |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2773186325000994 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Beta-glucan, a naturally occurring polysaccharide from sources such as oats, barley, yeast, and mushrooms, has gained significant attention for its health benefits, particularly in immune modulation and cholesterol reduction. However, a comprehensive understanding of the relationship between beta-glucan's molecular characteristics-such as chemical composition, molecular weight, and solubility and its biological effects remains limited. This study addresses this gap by employing advanced machine learning techniques, including regression models, random forests, support vector machines, and neural networks, to develop predictive models that link structural and physicochemical attributes of beta-glucan to its health benefits. The novelty of this research lies in its integrative computational approach, which goes beyond traditional experimental and simplistic statistical methods to uncover complex, non-linear structure-activity relationships. By providing a robust framework for analyzing diverse beta-glucan sources, the study offers actionable insights for the design of optimized formulations and targeted delivery systems, enhancing their efficacy in specific health applications. The Random Forest model achieved a prediction accuracy of 78 % (R² = 0.78) for cholesterol reduction, representing a ∼30 % improvement over baseline linear models. |
|---|---|
| ISSN: | 2773-1863 |