Antimicrobial resistance crisis: could artificial intelligence be the solution?

Abstract

Antimicrobial resistance is a global public health threat, and the World Health Organization (WHO) has announced a priority list of the most threatening pathogens against which novel antibiotics need to be developed. The discovery and introduction of novel antibiotics are time-consuming and expensive. According to WHO's report of antibacterial agents in clinical development, only 18 novel antibiotics have been approved since 2014. Therefore, novel antibiotics are critically needed. Artificial intelligence (AI) has been rapidly applied to drug development since its recent technical breakthrough and has dramatically improved the efficiency of the discovery of novel antibiotics. Here, we first summarized recently marketed novel antibiotics, and antibiotic candidates in clinical development. In addition, we systematically reviewed the involvement of AI in antibacterial drug development and utilization, including small molecules, antimicrobial peptides, phage therapy, essential oils, as well as resistance mechanism prediction, and antibiotic stewardship.

Keywords: Antibiotic; Antibiotic stewardship; Antimicrobial peptide; Artificial intelligence (AI); Clinical development; Machine learning (ML); Phage therapy.

Fig. 1

The pipeline (a) and the status (b) of novel antibiotic development. Novel drugs discovered in laboratories need to go through several stages including Investigational New Drug (IND) application, clinical development, and New Drug Application (NDA) before they become approved. Non-traditional chemicals include bacteriophage/phage products (n = 28), indirect-acting small molecules (n = 23), large molecules (n = 19), biologics (antibody or others, n = 8), immunomodulators (n = 7), nucleic acid-based products (n = 4), indirect-acting peptide (n = 2), and microbiome modifying agents (n = 1)

Fig. 2

Artificial intelligence (AI) in small-molecule antibiotic development. AI-based methods empower novel small-molecule antibiotic discovery from multiple dimensions, including mining secondary metabolites encoded by biosynthetic gene clusters (BGCs), screening existing compound libraries, and repurposing the Food and Drug Administration (FDA)-approved drugs. AI-based prediction of protein structures and functions, such as AlphaFold2 and RoseTTAFold, remarkably expands the protein space for docking simulations and drug rational design

Fig. 3

Artificial intelligence (AI) in the development of antimicrobial peptides (AMPs). AMPs databases have laid a solid foundation for AI-based model training, including natural language processing and deep generative networks. AI models can then be used to mine a wide range of protein sequence space, including the extinct human proteome, while high-throughput methods like cell-free synthesis significantly accelerate the speed of validation of candidate AMPs

Fig. 4

Artificial intelligence (AI) in the development of phage therapies. AI-based models have played a significant role in studying phages from their natural sources. This includes identifying phages from metagenomic samples, annotating phage virion proteins from phage genome sequences, predicting phage hosts, and determining phage lifestyles. These efforts lay a solid foundation for developing novel phage therapies

Fig. 5

Artificial intelligence (AI) in deciphering the mechanisms of action (MOA) and resistance mechanisms of novel antibacterials. Comparing cellular responses of bacteria before and after treatment of an antibacterial compound through multidimensional profiling enables AI-based methods to delineate the MOA of compounds and predict mechanisms of arising resistance. AMR antimicrobial resistance