Computational Drug Discovery in the Pandemic Era: AI-Enhanced Virtual Screening Approaches and Medical Informatics

Authors: Esther, Akinma Oluwabusayo; Akerele, Ifedayo David; Deborah, Ojo Ifeoluwa

Source journal: Healthcare informatics research (2026-10-01)

DOI: 10.4258/hir.2026.32.4.306

Abstract

OBJECTIVES: This narrative review examines how artificial intelligence-enhanced virtual screening (AI-VS), including multitarget screening strategies and hybrid AI-docking workflows, and the integration of medical informatics advance antiviral drug discovery and pandemic preparedness. METHODS: We reviewed recent studies that used graph neural networks to predict binding affinity, molecular dynamics simulations to validate binding stability, absorption, distribution, metabolism, excretion, and toxicity (ADMET) profiling frameworks, and protein structure prediction tools, including AlphaFold2 (2021) and AlphaFold3 (2024). A nonsystematic search of PubMed, Scopus, and Google Scholar identified seven representative studies published between 2020 and 2026, which were selected for their relevance, methodological innovation, and clinical applicability. RESULTS: Molecular dynamics simulations validated binding stability, while multi-target screening identified natural product inhibitors with favorable ADMET profiles. The hybrid AI-docking platform outperformed traditional docking, with VirtuDockDL achieving 99% accuracy and an area under the receiver operating characteristic curve of 0.99 on benchmark datasets, although external validation across protein families remains limited. These values apply to the benchmark dataset used and should not be generalized to other protein families, antiviral targets, or real-world drug discovery settings. CONCLUSIONS: AI-VS has the potential to shorten computational screening timelines from years to weeks and provide scalable frameworks that may support pandemic preparedness; this acceleration concerns in silico screening, not the full experimental or clinical development process. These methods illustrate how computational-first approaches can accelerate early therapeutic discovery, although they did not directly lead to the approval of nirmatrelvir or molnupiravir. Future priorities include expanding chemical probe libraries, improving the precision of scoring functions, and integrating AI-VS with medical informatics infrastructure.

Read the full text

Related topics

Related open access articles

Abstract and bibliographic data from PubMed (U.S. National Library of Medicine). Copyright of the abstract belongs to the original publisher or authors.