The abstract is read at the publisher; the summary is JClub's.
Surrogate-guided self-play Monte Carlo tree search produced 16 antimicrobial peptide candidates, with AMP2, AMP7, and AMP9 achieving complete inhibition of at least one strain (18.75%) in a 37-strain panel.
- Why it matters: De novo AMP design is hindered by a huge sequence space and the need to balance antimicrobial potency with toxicity, while surrogate scores may not reflect assay-specific activity.
- What they did: The framework jointly optimized predicted antimicrobial activity, HemoDL hemolytic propensity, and ESMFold structural confidence, then prioritized candidates for agar dilution testing across 37 strains at 0.25–64 μg/ml.
- The result: Only 3 of 16 peptides showed complete inhibition, while AMP12 and AMP13 reduced growth without full inhibition and 11 showed no complete inhibition, quantifying a gap between computational enrichment and experimental antibacterial outcomes.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.