OmegAMP preprint: generative AI designs programmable antimicrobial peptides, 204 experimentally validated with in vivo efficacy

AllThingsApx · x · 2026-09-16

A new bioRxiv preprint from the de la Fuente lab at UPenn, with Helmholtz Munich and TUM, introduces OmegAMP — a generative AI approach that moves from producing candidate molecules to designing peptides with controllable properties. The model controls fundamental physicochemical properties (length, charge, hydrophobicity) and supports analog and motif-guided design. The team experimentally characterized 204 peptides, identifying potent antimicrobials, multifunctional molecules, and lead designs with in vivo efficacy against A. baumannii. The broader goal: using machines not just to discover biological function but to engineer it for human health.

Original post →

More from Research

Research channel →