Researchers at the University of Pennsylvania introduced PepMLM, an AI-driven protein language model designed to develop peptide binders for challenging therapeutic targets lacking resolved structures. Trained on extensive peptide-protein sequence data, the platform outperforms existing methods in generating small, functional peptides capable of binding disease-related proteins implicated in conditions like Huntington’s disease and leukemia. Published in Nature Biotechnology, this technology advances drug discovery by targeting previously intractable proteins.