mMeta

mMeta performs multi-marker meta-analysis across multiple studies and related markers to estimate pooled effects, standard errors, and perform hypothesis testing while accounting for unknown marker-by-marker correlations.

Key Features:

• Multi-Marker Meta-Analysis (mMeta): Provides pooled estimates and standard errors across studies and markers and computes weighted averages of marker-specific pooled estimates for effect estimation.
• Non-parametric correlation estimation: Estimates marker-by-marker correlations non-parametrically using permutations.
• Adaptive Multi-Marker Meta-Analysis (aMeta): Performs hypothesis testing that maximizes power by using the minimum P-value among marker-specific meta-analyses.
• Core functions: Implements mMeta and aMeta via the mMeta.aMeta function and provides result plotting via mMeta.aMeta.plot.
• Example dataset: Includes HIV.Sum.Data as a summary dataset for demonstration in HIV-related microbiome analyses.

Scientific Applications:

• Microbiome meta-analysis: Synthesize multiple related markers such as α-diversity indices across studies to assess microbial diversity associations.
• HIV-associated diversity analyses: Compare α-diversity and other microbial markers between HIV-infected and uninfected groups across studies.
• General multi-marker synthesis: Integrate evidence across studies and correlated markers when marker-by-marker correlations are unknown.

Methodology:

Computes pooled estimates and standard errors via weighted averages of marker-specific pooled estimates, estimates marker-by-marker correlations non-parametrically using permutations, applies adaptive testing by taking the minimum P-value among marker-specific meta-analyses, and has been evaluated in silico and on real-world microbiome datasets.