APCyc: Property-Informed Design of Cyclic Peptides via Automated Cyclization

In three linesAPCyc is a target-aware de novo cyclic peptide generation framework that explicitly models cyclization and jointly optimizes multiple physicochemical properties. The model uses an expanded residue vocabulary and Bayesian posterior guidance to generate cyclization-aware peptides adapted to specific therapeutic targets.

Summary generated by Claude — human-verified