Tuberculosis remains the world's deadliest single-agent infection, killing 1.23 million people in 2024, per the WHO. Its tough outer membrane renders most antibiotics useless, creating an urgent need for novel treatments.
A University of Massachusetts Amherst team has created a pair of techniques to accelerate drug discovery. The approach combines artificial intelligence with a biological method nicknamed PAC-MAN, which rapidly screens compounds for their ability to penetrate Mtb's defenses.
Unlike traditional screening that tests each drug candidate in isolation, these methods allow researchers to test multiple possibilities in parallel. The techniques focus on identifying molecules that can breach the bacterium's notoriously stubborn cell membrane.
The breakthrough could dramatically shorten the timeline for finding new TB therapies. Current drug discovery for Mtb often takes years with high failure rates.
Further validation is needed before these techniques can be widely adopted. Outside researchers have not yet replicated the findings, and the methods have only been demonstrated in laboratory conditions.