The Easy Antibiotics Are Gone. Can Chemists Keep Up?
Science·October 26, 2026
For much of the twentieth century, antibiotic discovery was a matter of sampling the natural world. Researchers screened soil microbes and fungi for compounds that could kill bacteria, and the most accessible winners were found first. That low-hanging fruit is now largely gone, and the search has moved into harder territory where success is rare.
The odds are the core problem. A chemist may spend an entire career synthesizing and testing candidate molecules and still never produce one that clears every hurdle. Each promising compound has to survive laboratory testing, safety studies and clinical trials, and most fall away at some stage. Economics make it harder still. New antibiotics are often held back for use only as a last resort to slow the spread of resistance, which limits sales and makes the business case difficult for drug developers.
That is where newer tools are being tested. Artificial intelligence can screen vast libraries of virtual molecules and flag the ones most likely to act against bacteria, narrowing the field before any compound is made in the lab. Genetics offers a different route. By sequencing bacteria and mapping the genes that drive resistance, scientists can identify which biological targets are the most vulnerable and design drugs to hit them. Neither approach is a guaranteed fix. Computer predictions still need laboratory confirmation, and a molecule that looks promising on screen can fail once it meets a living body.
The second episode of Discovery's three-part series, reported by Roland Pease, examines whether these methods can shift the odds enough. The stakes are already concrete, since resistant infections are costing lives today. The question the episode raises is whether smarter chemistry can keep pace with bacteria that evolve around every new drug, or whether the pipeline stays too thin to matter.
Reporting based on an external source.