MNNMDA: Predicting human microbe-disease association via a method to minimize matrix nuclear norm

Area of Science:

• Microbiology
• Computational Biology
• Bioinformatics

Background:

• Identifying microbe-disease associations (MDAs) is crucial for understanding disease pathology.
• Traditional culture-based methods for MDAs are costly and time-consuming.
• Computational approaches are needed to prioritize potential MDAs for experimental validation.

Purpose of the Study:

• To propose a novel computational method, MNNMDA, for predicting microbe-disease associations.
• To leverage matrix nuclear norm minimization for accurate MDA prediction.
• To validate the effectiveness of MNNMDA using established datasets and compare it with existing methods.

Main Methods:

• Calculated Gaussian interaction profile kernel similarity and functional similarity for microbes and diseases.
• Constructed a heterogeneous information network integrating disease similarity, microbe similarity, and known MDAs.
• Formulated MDA prediction as a low-rank matrix completion problem solved via nuclear norm minimization.

Main Results:

• MNNMDA achieved high AUROC values (0.9536 on HMDAD, 0.9364 on Disbiome) outperforming state-of-the-art methods.
• Demonstrated robust performance across different dataset sizes and cross-validation strategies.
• Case studies on colon cancer and inflammatory bowel disease (IBD) confirmed MNNMDA's predictive power.

Conclusions:

• MNNMDA is an effective and accurate computational method for predicting microbe-disease associations.
• The approach aids in prioritizing microbes for further experimental investigation.
• MNNMDA contributes to advancing the understanding of microbe-associated diseases.