KMGTMDA: KAN-driven multiscale graph neural network and context-enhanced prediction for human microbe–disease associations

Key Points

• Improved prediction of human microbe-disease associations demonstrates significant accuracy gains.
• The KAN-driven multiscale graph neural network achieves 85% accuracy in predicting associations across diverse diseases.
• Analysis using advanced graph neural network techniques highlights complex relationships in microbiome data.
• The findings suggest potential applications in identifying microbial targets for disease prevention and treatment.