Compositional Data Analysis of Periodontal Disease Microbial Communities

Area of Science:

• Oral microbiome research
• Microbial ecology
• Host-microbiome interactions

Background:

• Periodontal disease (PD) is a chronic, polymicrobial condition with a significant host immune response.
• Next-generation sequencing (NGS) has advanced understanding of PD biodiversity and host specificity.
• Standard normalization methods in microbiome studies often overlook the compositional nature of data.

Purpose of the Study:

• To reanalyze existing periodontal multiomics data using compositional data analysis (CoDA) techniques.
• To identify novel microbial and host associations with periodontal disease.
• To evaluate the utility of CoDA in analyzing oral microbiome data.

Main Methods:

• Application of CoDA approaches, including centered log-ratio (clr) transformation, to multiomics data (16S, metagenomics, metabolomics).
• Reanalysis of previously published periodontal disease studies.
• Network analysis of microbial communities in relation to periodontal pocket depth.

Main Results:

• CoDA successfully identified novel associations, including the genera *Schwartzia* and *Aerococcus*, and C-reactive protein (CRP).
• Network analysis indicated reduced microbial connectivity in deeper periodontal pockets, suggesting a more random microbiome.
• Validated original findings while uncovering new insights into PD-associated features.

Conclusions:

• CoDA methods are effective for analyzing compositional multiomics data in periodontal research.
• This approach can reveal spurious correlations and identify novel disease-associated features.
• The study highlights the importance of appropriate statistical methods for microbiome data analysis in periodontology.