MPrESS

MPrESS estimates statistical power and required sample size for studies comparing microbiome taxon distributions using 16S rRNA gene composition data.

Key Features:

• Power Calculation: Predicts statistical power for detecting differences between populations using 16S rRNA gene composition data.
• Dirichlet Mixture Modeling (DMM) Simulation: Uses Dirichlet mixture modeling (DMM) to simulate microbiome 16S rRNA gene composition data that mirror observed diversity.
• Handling Variability: Quantifies variability in power estimates and reports that DMM simulations may overestimate power by less than 10%, with emphasis on highly discriminating taxa.
• Integration with Experimental Data: Allows mixing simulated and experimental 16S rRNA datasets to evaluate effects on power estimates.
• Sample Size Estimation: Computes sample size requirements to achieve specified statistical power levels for 16S rRNA microbiome comparisons.

Scientific Applications:

• Comparative microbiome study design: Designing 16S rRNA microbiome studies to detect population-level differences in microbial community composition.
• Power planning for limited samples: Estimating power and sample size when experimental datasets are small or when additional simulated samples are required.
• Microbiome association studies: Informing investigations of microbiome associations with health, disease, and environmental interactions through power-informed study design.

Methodology:

Performs Dirichlet mixture modeling (DMM) to simulate 16S rRNA gene composition data and applies power calculation and sample size estimation analyses on simulated and/or experimental datasets.