MicrobiomeGWAS: A Tool for Identifying Host Genetic Variants Associated with Microbiome Composition

Area of Science:

• Microbiome research
• Human genetics
• Statistical bioinformatics

Background:

• The human microbiome's composition is influenced by host genetics and environmental factors.
• Understanding these genetic influences is crucial for characterizing microbiome variation and improving genetic risk prediction.
• Microbiome community structure is best assessed using beta diversity metrics.

Purpose of the Study:

• To develop a statistical framework and software (microbiomeGWAS) for identifying host genetic variants associated with microbiome beta diversity.
• To account for interactions between host genetics and environmental factors.
• To provide accurate p-value approximations for association testing.

Main Methods:

• Developed a statistical framework and computationally efficient software package, microbiomeGWAS.
• Addressed positive skewness and kurtosis in score statistics for pairwise microbiome data.
• Developed accurate p-value approximations by correcting for skewness and kurtosis, validated through simulations.

Main Results:

• The developed methods provide accurate p-value approximations, overcoming limitations of asymptotic distributions.
• Application to lung tissue microbiome data demonstrated the effectiveness of skewness and kurtosis correction.
• Preliminary evidence suggests an association between six lung cancer risk SNPs and lung microbiome composition.

Conclusions:

• The microbiomeGWAS framework facilitates large-scale genome-wide association studies of the human microbiome.
• Accurate statistical methods are essential for robustly identifying genetic associations with microbiome beta diversity.
• This approach aids in understanding biological mechanisms and improving genetic risk prediction for diseases influenced by the microbiome.