Artificial Intelligence as a Catalyst for Antimicrobial Discovery: From Predictive Models to De Novo Design.

Area of Science:

• Computational biology and cheminformatics
• Drug discovery and development
• Infectious diseases and public health

Background:

• Antimicrobial resistance (AMR) is a critical global health threat, necessitating novel antibiotic discovery strategies.
• Traditional drug development is slow, expensive, and struggles against multidrug-resistant pathogens.
• Artificial intelligence (AI) offers transformative potential for accelerating antimicrobial discovery.

Purpose of the Study:

• To review and synthesize recent advancements in AI-driven approaches for discovering and designing small-molecule antibiotics and antimicrobial peptides.
• To critically assess the application of various AI techniques, including machine learning, deep learning, and generative models.
• To identify emerging trends and future directions in AI-enabled antimicrobial research.

Main Methods:

• Review of recent literature on AI applications in antimicrobial discovery.
• Analysis of machine learning, deep learning, and generative models (e.g., graph neural networks, transformers, VAEs, LLMs).
• Focus on AI applications in virtual screening, activity prediction, mechanism-informed prioritization, and de novo design.

Main Results:

• AI facilitates the identification of structurally novel antimicrobial compounds.
• AI enables the development of narrow-spectrum antimicrobials and improves peptide prediction interpretability.
• AI approaches are being used for virtual screening, activity prediction, and de novo design of antimicrobials.

Conclusions:

• AI significantly accelerates antimicrobial discovery and expands the accessible chemical and biological space.
• Challenges include data limitations, experimental validation needs, and clinical translation barriers.
• Future efforts should focus on integrating AI with experimental validation to bridge the gap between in silico discovery and therapeutic development.