AI identifies 13-species bacterial signature to diagnose gum disease
IBM Research, Salient Bio, and the Hartree Centre use machine learning to shift periodontitis diagnosis from single pathogens to microbial clusters.
Researchers have developed an AI-powered diagnostic approach that identifies periodontitis by analyzing complex interactions within the oral microbiome. The system moves beyond traditional methods by focusing on bacterial clusters rather than individual harmful species to detect gum disease.
The project, a collaboration between the Hartree National Centre for Digital Innovation (HNCDI), IBM Research, and Salient Bio, analyzed genetic data from approximately 600 oral microbiome samples. By applying machine learning to this data, the team distilled the vast microbial landscape into a specific 13-species diagnostic signature. This streamlined signature provides diagnostic accuracy comparable to models that utilize significantly larger and more complex datasets.
The shift to microbial clusters
Traditional diagnosis of periodontitis typically relies on the presence or absence of specific, known harmful bacteria. However, the researchers noted that gum disease is driven by a breakdown in the balance of the entire microbial community. The human mouth is the second most diverse microbial ecosystem in the body after the gut, containing over 700 species of bacteria, fungi, and viruses.
By integrating lifestyle and environmental data, the AI model identifies how these species interact in clusters. "Collaborating with the Hartree Centre and IBM Research helped us to bring together a great set of expertise to help us find association between bacteria in the microbiome and disease which has huge applications across the healthcare system," said Tom Sewell, Lead Bioinformatician at Salient Bio.
Implications for preventative care
This shift toward a cluster-based diagnostic allows for earlier intervention and more precise preventative care. Instead of relying on broad-spectrum antibiotics, clinicians can potentially employ targeted therapies based on the specific microbial signature of the patient. This precision is a critical step in combating global antibiotic resistance while managing a chronic condition that affects hundreds of millions of people worldwide.
Kate Royse, Director of the Hartree Centre, described the research as an example of AI's potential in preventative medicine, noting that the technology helps clinicians address disease before it takes hold rather than simply treating it after the fact.
Future applications
While the current focus is on periodontitis, the researchers suggest that this diagnostic framework is not limited to oral health. The methodology of using AI to identify specific microbial signatures within a diverse ecosystem could be applied to other chronic diseases linked to the microbiome. Future efforts will focus on validating these signatures across larger populations to refine the accuracy of early-stage detection.