Analysis of oral microbiome in glaucoma patients using machine learning prediction models

Purpose: The microbiome is considered an environmental factor that contributes to the progression of several neurodegenerative diseases. However, the association between microbiome and glaucoma remains unclear. This study investigated the features of the oral microbiome in patients with glaucoma and analyzed the microbiome biomarker candidates using a machine learning approach to predict the severity of glaucoma. Methods: The taxonomic composition of the oral microbiome was obtained using 16S rRNA gene sequencing, operational taxonomic unit analysis, and diversity analysis. The differentially expressed gene (DEG) analysis was performed to determine the taxonomic differences between the microbiomes of patients with glaucoma and the control participants. Multinomial logistic regression and association rule mining analysis using machine learning were performed to identify the microbiome biomarker related to glaucoma severity. Results: DEG analysis of the oral microbiome of patients with glaucoma revealed significant depletion of Lactococcus (P = 3.71e(−31)), whereas Faecalibacterium was enriched (P = 9.19e(−14)). The candidate rules generated from the oral microbiome, including Lactococcus, showed 96% accuracy for association with glaucoma. Conclusions: Our findings indicate microbiome biomarkers for glaucoma severity with high accuracy. The relatively low oral Lactococcus in the glaucoma population suggests that microbial dysbiosis could play an important role in the pathophysiology of glaucoma.