An international research team has developed an artificial intelligence tool named PepMLM capable of designing peptide-based drugs targeting “undruggable” proteins. The study, published in Nature Biotechnology, opens new avenues for treating challenging diseases such as cancer and neurodegenerative disorders.

The research was a collaboration between scientists from McMaster University, Duke University, and Cornell University. Unlike traditional drug design approaches, PepMLM does not rely on protein structural information—it designs targeting peptides solely from amino acid sequences. Lead researcher Pranam Chatterjee stated: “This breakthrough allows us to target disease proteins that lack stable structures.”

Experimental results show that peptides designed by PepMLM not only bind specifically to target proteins but, in some cases, can also promote the degradation of harmful proteins. Christina Peng from McMaster University noted: “In Huntington’s disease models, these AI-designed peptides demonstrated promising therapeutic effects.”

The research team is currently developing next-generation algorithms to further optimize the stability and targeting specificity of peptide drugs. Chatterjee added: “Our goal is to build a complete platform from sequence to drug development.”