Predicting resistance evolution to guide antibiotic development

Antibiotic resistance has emerged as a major bottleneck in antibiotic development, frequently undermining promising drug candidates and negating years of research and pharmaceutical investment. Despite growing recognition of this issue, discovery pipelines still emphasize potency and target specificity whereas resistance potential is often considered later in the development process. Predicting resistance remains challenging owing to the diversity of genetic mechanisms, species-specific adaptive pathways and potential side effects of resistance on bacterial viability. However, technological advances now enable systematic mapping of resistance evolution, the dissemination of resistance genes and prediction of the clinical impact. To improve early identification of resistance-prone antibiotic candidates, it is crucial to evaluate five key parameters, including de novo resistance evolvability across pathogens, resistance-fitness-virulence trade-offs, resistance stability, cross-resistance potential and health risk of resistance genes. Integrating these factors provides a systematic, resistance-based framework for classifying new antibiotic candidates and guiding compound prioritization and refinement. Utilizing this framework within the antibiotic development pipeline shifts resistance prediction from retrospective observation to a prospective guiding principle in drug design.