Solving the Computational Peptide Design Puzzle: Challenges and Opportunities

Abstract

Peptides are increasingly recognized as promising therapeutics, antimicrobial agents, and supramolecular biomaterials due to the remarkable versatility encoded in their amino acid sequences. However, the rational design of peptides remains challenging because sequence-structure–function relationships are not yet fully understood. Experimental exploration of peptide chemical space is inherently limited by the vast number of possible sequences and the complexity of resolving peptide behavior at molecular resolution. Computational methods, particularly molecular dynamics simulations and, more recently, machine learning and deep learning, have emerged as promising strategies to address these limitations. Although computational objectives differ, ranging from binding poses and biological activity to self-assembly propensity and supramolecular morphologies, they share common challenges associated with methodological heterogeneity, incomplete experimental metadata, insufficient benchmarking, and the lack of standardized and interoperable data sets. Here, we present our perspective on the main challenges in the computational study and design of peptides and discuss how collective efforts toward standardized data generation and data-driven discovery frameworks may help advance the field.