Predicting Metabolite-Disease Associations Based on Spy Strategy and ABC Algorithm.

Area of Science:

• Biomedical informatics
• Computational biology
• Metabolomics

Background:

• Metabolites are increasingly recognized for their correlation with complex human disease diagnosis.
• Accurate prediction of metabolite-disease associations is crucial for advancing biomedical research.
• Existing computational methods face challenges, particularly with incomplete association data.

Purpose of the Study:

• To propose a novel computational method, SSABCMDA, for predicting latent metabolite-disease associations.
• To enhance prediction accuracy by addressing missing associations and optimizing algorithm parameters.
• To validate the method's effectiveness through cross-validation and case studies.

Main Methods:

• Developed a novel method (SSABCMDA) integrating a spy strategy and an artificial bee colony (ABC) algorithm.
• Employed a spy strategy to extract reliable negative samples from unconfirmed metabolite-disease pairs.
• Utilized the ABC algorithm to optimize method parameters for improved performance.

Main Results:

• The SSABCMDA method demonstrated excellent predictive performance in cross-validation experiments.
• Case studies on three common diseases confirmed the method's validity and utility.
• Experimental results indicate effective prediction of potential metabolite-disease associations.

Conclusions:

• The proposed SSABCMDA method offers a robust approach for predicting metabolite-disease associations.
• This tool can effectively identify potential links, contributing to disease diagnosis and research.
• The integration of spy strategy and ABC algorithm optimization enhances prediction accuracy.