An empirical Bayes approach to normalization and differential abundance testing for microbiome data

Area of Science:

• Microbiome Research
• Bioinformatics
• Computational Biology

Background:

• Microbiome data analysis faces challenges like variable sequencing depth, sparsity, and compositional nature.
• Existing normalization methods for microbiome data have limitations.
• The phylogenetic structure of microbial communities is often overlooked in normalization.

Purpose of the Study:

• To develop a novel normalization method for microbiome data.
• To incorporate phylogenetic information into the normalization process.
• To improve differential abundance analysis in microbiome studies.

Main Methods:

• Proposed an empirical Bayes approach for microbiome data normalization under a multinomial distribution.
• Extended the method using a tree-based Dirichlet prior to integrate phylogenetic relationships.
• Developed a phylogeny-aware detection procedure for differential abundance testing.

Main Results:

• The proposed empirical Bayes normalization method demonstrates superior performance.
• Incorporating phylogenetic information enhances normalization accuracy.
• The phylogeny-aware detection procedure improves differential abundance analysis.

Conclusions:

• The empirical Bayes method offers a significant advancement in microbiome data normalization.
• This approach effectively addresses challenges in microbiome data analysis.
• The method shows superior performance compared to existing techniques in simulations and gut microbiome data.