Unfolding and De-confounding: Biologically meaningful causal inference from longitudinal multi-omic networks using...

Area of Science:

• Microbiome Research
• Systems Biology
• Bioinformatics

Background:

• Integrating multi-omic microbiome data is challenging for discovering microbial interactions.
• Existing causal inference tools struggle with complex longitudinal multi-omics datasets.

Approach:

• Developed METALICA, a novel suite of tools for inferring microbiome entity interactions.
• Introduced 'unrolling' to identify intermediary genes and metabolites.
• Implemented 'de-confounding' to detect spurious relationships caused by common factors.

Key Points:

• METALICA uncovers biologically meaningful multi-omic interactions.
• Identifies potential intermediaries (genes, metabolites) in microbial networks.
• Distinguishes direct interactions from spurious correlations.

Conclusions:

• METALICA enhances causal inference in longitudinal microbiome studies.
• Validated findings using literature and databases for IBD microbiome data.
• Supports predictions of microbial interactions and biological processes.