MECAF

MECAF performs differential abundance testing on high-dimensional microbiome compositional data to detect condition-associated changes without assuming equal covariance across groups.


Key Features:

  • Centered Log-Ratio Transformation: Employs centered log-ratio (CLR) transformation to analyze microbiome compositions and compare two high-dimensional abundance data matrices under different conditions.
  • Closed-Form P-value Computation: Computes test p-values directly using a closed-form solution derived from the asymptotic null distribution.
  • Maximum-Type Equal-Covariance-Assumption-Free Test: Implements a maximum-type test that operates without assuming equal covariance across groups.
  • Asymptotic Power Analysis Against Sparse Alternatives: Investigates asymptotic statistical power specifically against sparse alternatives common in microbiome studies.
  • Controlled Type I Error via Simulations: Demonstrates well-controlled type I error rates and improved statistical power relative to competing methods through simulation studies.
  • Compatibility with High-Throughput Metagenomic Sequencing: Applicable to abundance data derived from high-throughput metagenomic sequencing technologies.

Scientific Applications:

  • Differential Abundance Discovery: Detects shifts in microbiome abundances between disease and treatment conditions.
  • Clinical Pathogen and Infectious Disease Studies: Supports analyses for clinical pathogen diagnosis and studies of infectious disease using metagenomic abundance data.
  • Real-World Microbiome Data Analysis: Applied to real-world microbiome datasets to assess condition-associated microbial dynamics.

Methodology:

Computational steps explicitly include centered log-ratio transformation of compositional abundance matrices, a maximum-type equal-covariance-assumption-free test statistic, closed-form p-value derivation from the asymptotic null distribution, asymptotic power analysis against sparse alternatives, and evaluation via simulation studies.

Topics

Details

License:
GPL-3.0
Cost:
Free of charge
Tool Type:
command-line tool
Operating Systems:
Mac, Linux, Windows
Programming Languages:
R
Added:
2/8/2023
Last Updated:
11/24/2024

Operations

Publications

Li Z, Yu X, Guo H, Lee T, Hu J. A maximum-type microbial differential abundance test with application to high-dimensional microbiome data analyses. Frontiers in Cellular and Infection Microbiology. 2022;12. doi:10.3389/fcimb.2022.988717. PMID:36389165. PMCID:PMC9650337.