AI-Powered Biomimetic Peptide Design: Prof Nicholas Dunne Contributes to Cutting-Edge Drug Delivery Research
The research focuses on cell-penetrating peptides, which are crucial for delivering therapeutic molecules across cell membranes acting as "Trojan horses" for macromolecules that struggle with cellular entry and can trigger immune responses. These peptides, 5-30 amino acids long, can be synthetically created, derived from natural sources, or be chimeras, and are classified by properties like being cationic, hydrophobic, amphipathic, or intrinsically disordered.
Traditional CPP design involves costly, time-consuming rational design or "trial and error" approaches, often relying on modifications of known sequences, limiting the discovery of novel and more effective peptides. However, artificial intelligence has become a promising alternative, with AI-powered tools analysing peptide characteristics and predicting in vivo behavior. Advanced methods use supervised machine learning to predict cell penetration without extensive prior knowledge, reducing development costs and increasing successful drug delivery.
The future emphasises user-friendly deep learning and bioinformatic tools accessible to bioscientists, with web-accessible platforms playing a crucial role in this advancement. While no definitive rules exist, in silico and AI-driven tools allow for rapid screening and the potential to discover novel, more effective and less toxic CPPs.