Jiaxuan's AMP screening work has been published in Adv. Sci. (Link). This study introduces a mechanism-driven screening approach that leverages machine learning-based computational models to identify antimicrobial peptide sequences capable of targeting bacterial membranes and forming pores. Screening metaproteomes from poison frogs, African clawed frogs, and human skin led to the identification and validation of seven novel peptides. Each exhibited potent antimicrobial activity while demonstrating minimal hemolysis and cytotoxicity. Liposome leakage assays confirmed membrane disruption, with three peptides showing broad-spectrum efficacy against both Gram-positive and Gram-negative bacteria. Single-molecule experiments visualized peptide oligomerization on membranes, and electrophysiological measurements directly verified pore formation by the three broad-spectrum AMPs, suggesting a strong correlation between pore-forming capability and broad-spectrum activity. This research establishes a promising, mechanism-driven strategy for discovering new antimicrobial agents to combat antibiotic resistance. We extend our congratulations to Jiaxuan and all collaborators on this significant publication. Congratulations!
