A powerful microbial group association test based on the higher criticism analysis for sparse microbial association...

Area of Science:

• Microbiome research
• Statistical genetics
• Computational biology

Background:

• Evaluating microbial group composition and host phenotype associations is crucial in human microbiome studies.
• Existing microbial association tests often struggle with sparse data, where signals are weak or infrequent.
• High-dimensionality, compositionality, and phylogenetic relationships are key features of microbiome data that pose analytical challenges.

Purpose of the Study:

• To develop a powerful statistical test for identifying microbial group associations in the presence of sparse signals.
• To create a data-driven omnibus test that adapts to varying levels of phylogenetic relevance and sparsity.
• To provide a robust analytical tool for microbiome association studies.

Main Methods:

• Introduced microbiome higher criticism analysis (MiHC), a novel omnibus test.
• Tailored the higher criticism test to incorporate phylogenetic information and modulate sparsity levels.
• Integrated the Simes test to address excessively high sparsity levels, enhancing adaptability.

Main Results:

• MiHC demonstrated high statistical power across different phylogenetic relevance and sparsity levels in simulations.
• The method maintained correct type I error rates, ensuring reliable findings.
• Applied MiHC to four real-world microbiome datasets, analyzing associations with smoking, delivery mode, type 1 diabetes, and HIV status.

Conclusions:

• MiHC is a valuable analytical tool due to its adaptivity to unknown phylogenetic relevance and sparsity patterns.
• The method robustly identifies microbial group associations in diverse microbiome datasets.
• MiHC is available as an R software package, facilitating its use in the research community.