New AI Tool Reveals How Drugs Kill Tuberculosis

---

Tuberculosis (TB) remains the world’s deadliest infectious disease and a major treatment challenge, especially in cases resistant to standard therapies. A new AI-powered tool developed by Tufts University researchers, called DECIPHAER (Decoding Cross-Modal Information of Pharmacologies via Autoencoders), seeks to advance TB treatment by elucidating exactly how drugs kill the bacterium at the cellular and molecular levels.By combining high-resolution imaging of TB bacteria during drug treatment (capturing morphological changes like altered cell shape or internal structure) with gene expression data, DECIPHAER uses machine learning to link visual "signatures" of cellular damage to specific molecular pathways of action. In a proof of concept, the tool revealed that one drug in development, previously assumed to kill TB by disrupting the bacterial cell wall, actually works by impairing the bacterium’s respiratory chain and energy production. Because DECIPHAER can infer a drug’s mechanism of death from images alone—rather than relying solely on expensive and time-consuming molecular assays—this method has promise to accelerate discovery of effective drug combinations. It may help shorten TB treatment regimens, improve efficacy against drug-resistant strains, and potentially be adapted for other infectious diseases and cancer.

Author:

Genevieve Rajewski

Audiences:

Educators, Outreach Workers, Physicians, Researchers

Topic:

TB Treatment

Language:

English