PepComLibGen: A Web Server for Generating Peptide Libraries for Computer-Aided De Novo Peptide Design and Combinatorial Lead Optimization

Abstract

Small peptides offer unique advantages for drug discovery due to their high specificity, low cost, and ability to disrupt protein-protein interactions, making them a promising chemical space for discovering new drugs and biotechnological tools. However, there is currently no available resource to generate a machine-readable library of small peptides from all possible sequences of canonical amino acids that can be used in a computer-aided drug discovery (CADD) campaign. To address this gap, we present a novel web server (www.peptidelibgen.com) that generates comprehensive SMILES libraries of small peptides from user-defined canonical amino acids and their cyclic analogues (including N-C terminal, side chain-N terminal, side chain-C terminal, side chain-side chain, and disulfide bridge). The server incorporates enantiomers of canonical amino acids, a user-specific noncanonical amino acid, end-capping, and library filters based on ease of synthesis, stability, and aqueous solubility. The generated peptides are output as smiles strings with structured FASTA identifiers, facilitating easy interpretation of top hits in a CADD workflow. To supplement this tool, we have also included an additional widget that allows one to rapidly obtain the SMILES codes for large libraries of combinatorial structures, especially useful for incorporation as unnatural residues when one seeks to screen a larger structural space. Overall, our web server provides a valuable resource for the discovery of new drug candidates using small peptides.