NetNiche: Microbe-Metabolite Network Reconstruction and Microbial Niche Analysis

Area of Science:

• Microbiology
• Bioinformatics
• Computational Biology

Background:

• Metagenomics and metabolomics are key for studying in vivo microbe-metabolite interactions.
• Accurate computational methods for inferring these interactions are currently lacking.

Purpose of the Study:

• To present NetNiche, a novel context-aware framework for graph representation learning.
• To accurately predict microbe-metabolite and microbe-microbe interactions by integrating abundance data and prior knowledge.

Main Methods:

• Developed NetNiche, a graph representation learning framework.
• Integrated microbe and metabolite abundance data with prior knowledge.
• Applied NetNiche to gut and soil microbiome datasets.

Main Results:

• NetNiche demonstrated superior performance compared to state-of-the-art methods like SPIEC-EASI, SparCC, and mmvec.
• Successfully predicted microbe-metabolite and microbe-microbe interactions in diverse microbiome datasets.

Conclusions:

• NetNiche is an effective computational tool for predicting microbe-metabolite and microbe-microbe interactions.
• The framework shows wide applicability in multi-omics studies, particularly for the human microbiome.