Phylogeny-guided microbiome OTU-specific association test (POST).

Area of Science:

• Microbiome research
• Computational biology
• Statistical genetics

Background:

• Host conditions and microbiome profiles (operational taxonomic units - OTUs) are crucial for understanding human health.
• High dimensionality and sparsity of microbiome data challenge traditional association testing.
• Phylogenetic information is often used, assuming similar taxa behave similarly, but this assumption may not always hold.

Purpose of the Study:

• To propose a novel statistical test, the phylogeny-guided microbiome OTU-specific association test (POST), for analyzing microbiome data.
• To develop a data-supervised approach for incorporating phylogenetic information in association testing.
• To extend existing kernel machine microbiome tests to the OTU level.

Main Methods:

• Introduced POST, a local collapsing test that adaptively borrows information from neighboring OTUs based on phylogenetic distance and outcome association.
• Utilized a kernel machine framework to handle complex OTU effects.
• Extended community-level microbiome tests to the OTU level.

Main Results:

• POST demonstrates superior performance over existing OTU-level tests when phylogenetic information is informative.
• POST achieves comparable performance to existing methods when phylogenetic information is not informative.
• Real-data applications identified biologically relevant, outcome-associated OTUs, similar to or exceeding existing methods.

Conclusions:

• Adaptively leveraging phylogenetic information with POST enhances the selection performance of associated microbiome features.
• POST improves both true-positive and false-positive detection rates.
• A user-friendly R package, POSTm, is available for public use.