Salivary Microbial Meta-Analysis for Multi-Disease Prediction, 7,750 Samples

7,750 saliva samples from 22 cohorts were analyzed to explore links between oral microbiota and disease. A meta-analysis of public 16S rRNA data identified nine core microbes and built a multi-disease prediction model. The study, authored by Qixiang Yuan and updated in April 2026, validated the feasibility of using saliva microbiota and machine learning for non-invasive diagnostics.

Use Cases

• Training multi-class random forest models for disease prediction based on genus-level microbial abundance.
• Identifying core salivary microbiota in healthy populations based on negative control group data.
• Comparing microbiota structure across physiological and pathological states, such as colorectal cancer or HIV.
• Establishing healthy baselines for oral microbiota to serve as a reference for disease studies.

Strengths

• Includes 7,750 samples from 22 cohorts, providing a substantial sample size.
• Reports robust model performance with AUC scores ranging from 0.898 to 1.0.
• Identifies nine specific core microbiota genera, such as g:Streptococcus.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect temporal bias inherent to the source, covering studies from 2016 to 2024.

Provenance

Source

figshare

Collection Method

Meta-analysis of public 16S rRNA data from PubMed (2016–2024), processed with QIIME2 and Wekemo.

Time Range

2016–2024

Freshness

Last updated 2026-04-15 05:46:11; freshness should be verified.