Identification of Active Components for Sports Supplements: Machine Learning-Driven Classification and Cell-Based...

Area of Science:

• Sports Science
• Pharmacology
• Biotechnology

Background:

• Identifying active components in herbal medicines is crucial for sports supplement development.
• High-throughput screening of these components presents a significant challenge.
• Machine learning offers a potential solution for efficient screening.

Purpose of the Study:

• To develop and validate machine learning models for screening active components in herbal medicines as potential sports supplements.
• To assess the efficacy of screened components using cell-based assays.
• To investigate the mechanism of action for promising compounds like luteolin.

Main Methods:

• Construction of six machine learning prediction models with accuracy >0.88.
• Screening of active components from herbal medicines using the developed models.
• Validation of screened components' potency on C2C12 cells.
• Investigation of luteolin's effects on skeletal muscle performance via immunofluorescence and high-content imaging.

Main Results:

• The machine learning models demonstrated high accuracy in screening active components.
• Eleven out of twelve randomly selected components significantly increased myotube diameters and protein synthesis in C2C12 cells.
• Luteolin was identified to enhance skeletal muscle performance by activating PGC-1α and MAPK signaling pathways.

Conclusions:

• High-throughput prediction models are effective tools for screening active components from herbal medicines for sports supplement applications.
• The study validates the use of machine learning in identifying novel sports supplement ingredients.
• Luteolin shows potential as a sports supplement ingredient due to its positive effects on skeletal muscle.