A generative deep learning approach to de novo antibiotic design

Clinical Snapshot

Key Findings

• P Value: Not reported in abstract

• Effect Size: 29% success rate for selective antibacterial activity

• Primary Outcome: 7 of 24 synthesized compounds demonstrated selective antibacterial activity

• Nnt Or Sensitivity: Two lead compounds showed bactericidal efficacy against MDR isolates with distinct mechanisms of action

• Confidence Interval: Not reported in abstract

Clinical Application

Early-stage research requiring extensive clinical development; computational platform could accelerate future antibiotic discovery Highly relevant to Australian AMR strategy; could support TGA fast-track pathways for novel antibiotics addressing critical resistance patterns prevalent in Australian healthcare settings Patients with infections caused by multidrug-resistant N. gonorrhoeae and S. aureus, particularly relevant for hospital-acquired and sexually transmitted infections

Abstract

The antimicrobial resistance crisis necessitates structurally distinct antibiotics. While deep learning approaches can identify antibacterial compounds from existing libraries, structural novelty remains limited. Here, we developed a generative artificial intelligence framework for designing de novo antibiotics through two approaches: a fragment-based method to comprehensively screen >107 chemical fragments in silico against Neisseria gonorrhoeae or Staphylococcus aureus, subsequently expanding promising fragments, and an unconstrained de novo compound generation, each using genetic algorithms and variational autoencoders. Of 24 synthesized compounds, seven demonstrated selective antibacterial activity. Two lead compounds exhibited bactericidal efficacy against multidrug-resistant isolates with distinct mechanisms of action and reduced bacterial burden in vivo in mouse models of N. gonorrhoeae vaginal infection and methicillin-resistant S. aureus skin infection. We further validated structural analogs for both compound classes as antibacterial. Our approach enables the generative deep-learning-guided design of de novo antibiotics, providing a platform for mapping uncharted regions of chemical space.

References

1. 1.Krishnan, A., Anahtar, M. N., Valeri, J. A., Jin, W., Donghia, N. M., Sieben, L., Luttens, A., Zhang, Y., Modaresi, S. M., Hennes, A., Fromer, J., Bandyopadhyay, P., Chen, J. C., Rehman, D., Desai, R., Edwards, P., Lach, R. S., Aschtgen, M.-S., Gaborieau, M., Gaetani, M., Palace, S. G., Omori, S., Khonde, L., Moroz, Y. S., Blough, B., Jin, C., Loh, E., Grad, Y. H., Saei, A. A., Coley, C. W., Wong, F., & Collins, J. J. (2025). A generative deep learning approach to de novo antibiotic design. Cell. https://doi.org/10.1038/s41592-022-01639-4