Accelerated antimicrobial discovery via deep generative models and molecular dynamics simulations

Abstract

The de novo design of antimicrobial therapeutics involves the exploration of a vast chemical repertoire to find compounds with broad-spectrum potency and low toxicity. Here, we report an efficient computational method for the generation of antimicrobials with desired attributes. The method leverages guidance from classifiers trained on an informative latent space of molecules modelled using a deep generative autoencoder, and screens the generated molecules using deep-learning classifiers as well as physicochemical features derived from high-throughput molecular dynamics simulations. Within 48 days, we identified, synthesized and experimentally tested 20 candidate antimicrobial peptides, of which two displayed high potency against diverse Gram-positive and Gram-negative pathogens (including multidrug-resistant Klebsiella pneumoniae) and a low propensity to induce drug resistance in Escherichia coli. Both peptides have low toxicity, as validated in vitro and in mice. We also show using live-cell confocal imaging that the bactericidal mode of action of the peptides involves the formation of membrane pores. The combination of deep learning and molecular dynamics may accelerate the discovery of potent and selective broad-spectrum antimicrobials.

Machine learning designs non-hemolytic antimicrobial peptides

“…3 Several ML models for AMP de novo design have been reported so far, and they range from classifiers for AMPs prediction applied to select sequences from randomly generated, existing, or genome derived libraries, [25][26][27][28][29][30][31][32][33] to standalone generative models, 34 to a combination of both generative models and classifiers. [35][36][37] Furthermore, ML has also been used in combination with evolutionary algorithms for the optimization of AMPs. 38,39 However, only two of the discussed studies considered both activity and hemolysis in the design of novel AMPs, 31,37 reflecting the challenge of avoiding hemolysis in designing AMPs, and highlighting the importance of its further investigation.…”

“…[35][36][37] Furthermore, ML has also been used in combination with evolutionary algorithms for the optimization of AMPs. 38,39 However, only two of the discussed studies considered both activity and hemolysis in the design of novel AMPs, 31,37 reflecting the challenge of avoiding hemolysis in designing AMPs, and highlighting the importance of its further investigation.…”

Nanoantibiotics: Functions and Properties at the Nanoscale to Combat Antibiotic Resistance

“…Biological and life sciences are advancing with the revolution of nanotechnology as a Frontier. Data mining, deep learning, and artificial intelligence technologies can be integrated to discover more antibiotic molecules (Stokes et al, 2020;Das et al, 2021) and their potential synergy with NPs to enrich the existing arsenal of antimicrobials.…”

Machine Learning Designs Non-Hemolytic Antimicrobial Peptides

“…A related yet significantly different study was recently published in which AMPs were designed using a ML autoencoder architecture exploiting activity and hemolysis data combined with molecular dynamics simulations to select α-helical sequences. 39…”