RAIDA

RAIDA (Ratio Approach for Identifying Differential Abundance): Ratio-based differential abundance analysis in metagenomics

RAIDA identifies differentially abundant features in metagenomic samples across conditions using a ratio-based statistical framework that models feature relationships to improve differential abundance detection.

Key Features:

• Ratio-Based Modeling: Utilizes ratios between features within a modified zero-inflated lognormal model to mitigate scale variability in feature counts within and across conditions.
• Normalization Robustness: Avoids bias from total sum, mean, or median normalization and remains unaffected by differences in total abundances of differentially abundant features (DAFs) across conditions.
• Simulation-Based Evaluation: Demonstrates consistent statistical power and improved performance relative to existing methods under specific simulated conditions.

Scientific Applications:

• Microbial Community Analysis: Detects differential abundance of microbial species or genes in microbial ecology, medical microbiology, environmental science, and biotechnology, including studies of type II diabetes.

Methodology:

RAIDA applies a ratio-based statistical framework by modeling pairwise feature ratios under a modified zero-inflated lognormal model to control for compositional effects and variability in total feature counts, enabling robust identification of differentially abundant features across experimental conditions.