StaPep: An Open-Source\nToolkit for Structure Prediction,\nFeature Extraction, and Rational Design of Hydrocarbon-Stapled Peptides

All-hydrocarbon stapled peptides, with their covalent\nside-chain\nconstraints, provide enhanced proteolytic stability and membrane permeability,\nmaking them superior to linear peptides. However, tools for extracting\nstructural and physicochemical descriptors to predict the properties\nof hydrocarbon-stapled peptides are lacking. To address this, we present\nStaPep, a Python-based toolkit for generating 3D structures and calculating\n21 features for hydrocarbon-stapled peptides. StaPep supports peptides\ncontaining two non-standard amino acids (norleucine and 2-aminoisobutyric\nacid) and six non-natural anchoring residues (S3, S5, S8, R3, R5,\nand R8), with customization options for other non-standard amino acids.\nWe showcase StaPep’s utility through three case studies. The\nfirst generates 3D structures of these peptides with a mean RMSD of\n1.62 ± 0.86, offering essential structural insights for drug\ndesign and biological activity prediction. The second develops machine\nlearning models based on calculated molecular features to differentiate\nbetween membrane-permeable and non-permeable stapled peptides, achieving\nan AUC of 0.93. The third constructs regression models to predict\nthe antimicrobial activity of stapled peptides against <i>Escherichia\ncoli</i>, with a Pearson correlation of 0.84. StaPep’s\npipeline spans data retrieval, structure generation, feature calculation,\nand machine learning modeling for hydrocarbon-stapled peptides. The\nsource codes and data set are freely available on Github: https://github.com/dahuilangda/stapep_package.