Tulane University researchers have made a significant breakthrough in the global fight against drug-resistant infections particularly those caused by deadly bacteria like *Mycobacterium tuberculosis* and *Staphylococcus aureus*. These infections are increasingly difficult to treat, often requiring more toxic or costly medications, and are associated with longer hospital stays and increased mortality. For instance, in 2021 alone, 450,000 people were diagnosed with multidrug-resistant tuberculosis, with treatment success rates dropping to just 57%, according to the World Health Organization.

In response to this urgent health challenge, Tulane scientists have developed an innovative AI-powered method that enhances the accuracy of detecting antibiotic resistance. Their new model, known as the Group Association Model (GAM), applies machine learning to identify genetic mutations linked to resistance without relying on previously known mechanisms. This allows the system to detect novel and rare genetic changes that traditional tools might miss.

Unlike conventional methods, which can take days or weeks (as with culture-based testing) or may fail to detect rare mutations (as with standard DNA-based diagnostics), Tulane’s GAM model analyzes the full genome of bacterial strains. By comparing resistant and non-resistant strains at the genomic level, it accurately pinpoints the mutations responsible for resistance like uncovering a bacterium’s genetic fingerprint to understand its antibiotic evasion tactics.

In their recent study published in *Nature Communications*, the researchers applied GAM to more than 7,000 strains of tuberculosis and nearly 4,000 strains of staph bacteria. The model outperformed existing methods, including the WHO’s resistance database, by identifying relevant mutations with higher precision and reducing false positives. This is critical, as false positives can lead to misdiagnoses and inappropriate treatments.

Moreover, when paired with machine learning, the model proved even more effective, especially when working with limited or incomplete data. In clinical samples from China, GAM predicted resistance to frontline antibiotics more accurately than traditional WHO-guided tests. This enables physicians to select targeted therapies earlier, potentially preventing infections from worsening or spreading.

What’s particularly promising is that this model doesn’t require prior knowledge of resistance pathways, making it adaptable across different bacteria or even in agricultural contexts where antibiotic resistance also poses a major concern.

By integrating cutting-edge AI with molecular data, this research marks a pivotal step toward the goals of molecular diagnostics and precision medicine. It empowers clinicians with the tools to personalize treatment strategies based on real-time, genomic insights an approach that could transform how we respond to antibiotic resistance in the future.

Source: https://news.tulane.edu/pr/tulane-researchers-use-ai-improve-diagnosis-drug-resistant-infections