A collaboration between STFC's Hartree Centre, IBM Research, and Salient Bio has led to the development of an AI-powered method for diagnosing periodontitis, a severe form of gum disease affecting hundreds of millions globally. This new approach moves beyond simply checking for specific harmful bacteria, instead focusing on the balance of the entire microbial community in the mouth.
The team analysed genetic data from mouth bacteria across 600 samples, mapping interactions between different species. Using machine learning, they identified bacterial clusters linked to both healthy and diseased mouths. From hundreds of variables, AI distilled the complexity to just 13 key bacterial species, which proved to be an accurate diagnostic panel.
This research also incorporated patient lifestyle and environmental data to build a more comprehensive picture of risk factors. The aim is to identify individuals at risk before symptoms become severe, allowing clinicians to intervene earlier with targeted therapies, potentially reducing the reliance on broad-spectrum antibiotics.
Tom Sewell, Lead Bioinformatician at Salient Bio, stated that the collaboration enabled the team to find associations between bacteria in the microbiome and disease, with applications across the healthcare system. The framework developed could also potentially be used to understand other chronic conditions linked to the human microbiome.