A two-stage microbial association mapping framework with advanced FDR control.

Area of Science:

• Microbiology
• Bioinformatics
• Statistical Genetics

Background:

• Detecting microbe-pathogen associations at specific taxonomic ranks (e.g., genus, species) is crucial in microbiome research.
• Traditional methods often neglect the evolutionary relationships within the taxonomic tree, potentially leading to reduced statistical power and conservative results.
• Leveraging the hierarchical structure of microbial taxa can enhance the power of association tests.

Purpose of the Study:

• To introduce massMap, a novel two-stage microbial association mapping framework.
• To enhance statistical power in identifying microbe-pathogen associations by utilizing the taxonomic tree structure.
• To improve upon traditional methods that individually test taxa and may overlook complex dependencies.

Main Methods:

• Propose a two-stage framework (massMap) that leverages taxonomic tree information for association mapping.
• Stage 1: Screen associations at a higher taxonomic rank using the OMiAT group test.
• Stage 2: Test associations for individual taxa at the target rank within significant groups, employing Hierarchical Benjamini-Hochberg (HBH) or selected subset testing (SST) for False Discovery Rate (FDR) control.

Main Results:

• Simulations demonstrate that massMap significantly alleviates the multiplicity issue and offers greater statistical power compared to traditional methods.
• massMap effectively controls the FDR at desired levels across various scenarios.
• Real data analyses show massMap identifies an equal or greater number of associated species with smaller adjusted p-values, confirming its efficiency.

Conclusions:

• massMap is an innovative framework for microbial association studies.
• It enhances efficiency by effectively utilizing the inherent taxonomic structure of microbiome data.
• The R package for massMap is publicly available for broader application.