Artificial Intelligence–Assisted In-Silico Design of Probiotic Oral Formulations for Microbiome Modulation and Caries Prevention

Preventive dentistry is undergoing a paradigm shift from antimicrobial eradication to microbiome modulation, emphasizing symbiotic oral ecology. Recent advances in machine learning (ML) and computational microbiology enable predictive modeling of probiotic–pathogen interactions, paving the way for precision-designed probiotic formulations. This research presents an AI-assisted in-silico framework for identifying probiotic consortia capable of suppressing cariogenic biofilms and restoring microbial equilibrium within the oral cavity.

A dataset comprising 420 metagenomic oral samples from children and adults was analyzed using graph-based microbial co-occurrence networks. Machine learning algorithms—including Random Forest, Gradient Boosting, and Neural Network classifiers—were trained to predict microbial dominance patterns under varying pH and sugar-exposure conditions. Candidate probiotic strains were virtually screened from the Human Oral Microbiome Database (HOMD) and assessed for competitive exclusion potential against Streptococcus mutans, Lactobacillus fermentum, and Prevotella intermedia.

Results revealed a synergistic tri-strain formulation (Lactobacillus rhamnosus, Bifidobacterium breve, Streptococcus salivarius) that achieved a 72% predicted suppression rate of cariogenic biofilm formation in simulation. Metabolic flux modeling confirmed elevated production of alkaline metabolites, enhancing biofilm pH buffering capacity. In-silico simulations also indicated enhanced resilience of the probiotic network under carbohydrate stress and acidogenic fluctuations.

The study demonstrates that AI-assisted microbiome modeling can rationally design probiotic interventions tailored to specific ecological niches of the oral cavity. This computational approach reduces experimental dependency, accelerates product development, and aligns with sustainable preventive dentistry. Future work will integrate wet-lab validation and personalized probiotic formulations driven by salivary microbiome sequencing.