Therapeutic peptide informatics in the artificial intelligence era: From curated resources to generative design

Therapeutic peptides are emerging as a powerful class of pharmaceuticals that combine the synthetic accessibility of small molecules with the specificity of biologics, yet poor stability and low oral bioavailability remain barriers. This review surveys an expanded, scope-defined set of 39 therapeutic peptide databases and 62 computational tools for activity prediction, physicochemical profiling, structural modeling and rational design. We highlight how machine learning and deep learning are accelerating data-driven peptide drug discovery and discuss the urgent need for standardized data frameworks and integrated evaluation platforms to bridge the gap between computational prediction and experimental validation.