AI-Driven Pipeline Identifies New Tuberculosis Drug Targets with Potential Impact on Treatment
Researchers have developed an AI-based screening pipeline that integrates machine learning and physics-based simulations to identify inhibitors of the mycobacterial toxin MenT3, a new target in tuberculosis (TB). This study, published in the journal Nature, utilized a combination of ADMET-AI, PharmacoNet, docking, and molecular dynamics simulations to filter over 100,000 compounds. Five promising candidates were identified that bind stably to MenT3, suggesting potential as next-generation anti-TB agents. The MenT3 toxin is part of a bacterial toxin-antitoxin system that regulates stress responses and promotes bacterial persistence within hosts. The study's approach combines generative machine learning with pharmacokinetic filtering and molecular docking to identify novel inhibitors, marking a significant advancement in TB drug discovery. 
