Predicting lifespan-extending chemical compounds for C. elegans with machine learning and biologically interpretable...

Area of Science:

• Gerontology and computational biology
• Application of machine learning in drug discovery

Background:

• Growing interest in pharmacological interventions for aging.
• Machine learning is increasingly used to analyze aging-related data.

Purpose of the Study:

• To use machine learning to analyze the DrugAge database for compounds that extend lifespan.
• To predict novel life-extending compounds using biological features.

Main Methods:

• Created four datasets based on compound-protein interactions, gene ontology terms, and phenotype ontology terms.
• Employed feature selection and the random forest algorithm for predictive modeling.
• Interpreted key features related to aging biology.

Main Results:

• Identified "Glutathione metabolic process" (a Gene Ontology term) as important for lifespan extension.
• Predicted nitroprusside, an antihypertensive drug, as a promising novel compound for extending lifespan.
• Developed predictive models for lifespan extension in *C. elegans*.

Conclusions:

• Machine learning can effectively predict novel compounds that extend lifespan.
• This approach opens new avenues for discovering anti-aging drugs.
• Biological features like glutathione metabolism are crucial targets for longevity interventions.