Abstract
Peptide-based therapeutics have emerged as a significant treatment strategy, offering high specificity and tunable pharmacokinetics. Recent advances in Artificial Intelligence (AI) have shifted the focus towards structure prediction, generative design, and interaction modelling, significantly accelerating drug design and discovery. Deep learning architectures, including graph neural networks, transformers, and diffusion models, have facilitated the generation of novel sequences for the target of interest, although predicting the solubility, immunogenicity, and toxicity of these sequences remains a challenge. Innovations in peptide chemistry, such as cyclization, stapling, non-canonical amino acids, and nanoparticle formulations, help overcome the hurdles of bioavailability and permeation. These chemical approaches, combined with developments in autonomous peptide synthesis and high-throughput screening, have considerably reduced discovery timelines from years to months. Clinically, this progress is apparent in the growing number of approved peptide drugs for metabolic disorders, oncology, and medical imaging. Here, we review recent progress in peptide-based drug design using AI, focusing on generative architectures and interactions. We then examine AI-driven screening and delivery optimization for these peptide-based discoveries. Finally, we discuss the current limitations, practical challenges, and future direction with particular emphasis on data quality and autonomous drug discovery.