Mulat Mulugeta, Banicod Riza Jane S, Tabassum Nazia, Javaid Aqib, Kim Tae-Hee, Kim Young-Mog, Khan Fazlurrahman
Department of Biotechnology, School of Bioscience and Technology, College of Natural Sciences, Wollo University, Dessie, Ethiopia.
Fisheries Postharvest Research and Development Division, National Fisheries Research and Development Institute, Quezon City 1103, Philippines.
J Microbiol Methods. 2025 Aug 20;237:107232. doi: 10.1016/j.mimet.2025.107232.
Artificial intelligence (AI) is revolutionizing antimicrobial drug discovery by delivering major improvements in precision, innovation, and efficiency for combating bacterial, fungal, and viral pathogens. Traditional approaches to developing treatments for microbial infections are often hampered by high costs, lengthy timelines, and frequent failures. Modern AI technologies, particularly deep learning, machine learning, computational biology, and big data analytics, provide robust solutions to these challenges by analyzing large-scale biological datasets to predict molecular interactions, identify promising treatment candidates, and expedite both preclinical and clinical development. Innovative techniques such as generative adversarial networks for novel compound discovery, reinforcement learning for optimizing antimicrobial candidates, and natural language processing for extracting knowledge from biomedical literature are now vital to infectious disease research. These approaches facilitate early toxicity prediction, microbial target identification, virtual screening, and the development of more individualized therapies. Notwithstanding these advances, challenges remain, including inconsistent data quality, limited interpretability, and unresolved ethical or legal concerns. This review examines recent advancements in AI applications for microbial drug discovery, with a focus on de novo molecular design, ligand- and structure-based screening, and AI-enabled biomarker identification. Remaining application barriers and promising future directions in AI-driven antimicrobial drug development are also elucidated. Collectively, these innovations are poised to accelerate the discovery of new therapies, reduce costs, and enhance patient outcomes in the fight against infectious diseases.
人工智能（AI）正在彻底改变抗菌药物的发现方式，在对抗细菌、真菌和病毒病原体方面，它在精准度、创新性和效率上都有了重大提升。传统的微生物感染治疗方法往往受到高成本、长周期和频繁失败的阻碍。现代人工智能技术，特别是深度学习、机器学习、计算生物学和大数据分析，通过分析大规模生物数据集来预测分子相互作用、识别有潜力的治疗候选物，并加快临床前和临床开发进程，为这些挑战提供了强有力的解决方案。创新技术，如用于新型化合物发现的生成对抗网络、用于优化抗菌候选物的强化学习以及用于从生物医学文献中提取知识的自然语言处理，现在对传染病研究至关重要。这些方法有助于早期毒性预测、微生物靶点识别、虚拟筛选以及更个性化疗法的开发。尽管取得了这些进展，但挑战依然存在，包括数据质量不一致、可解释性有限以及未解决的伦理或法律问题。本综述探讨了人工智能在微生物药物发现应用中的最新进展，重点关注从头分子设计、基于配体和结构的筛选以及人工智能驱动的生物标志物识别。还阐明了人工智能驱动的抗菌药物开发中剩余的应用障碍和有前景的未来方向。总体而言，这些创新有望加速新疗法的发现，降低成本，并在抗击传染病中改善患者预后。
