AI-Driven Antiviral Drug Discovery and Repurposing: From Prediction to Experimental Validation

The increasing frequency of emerging and re-emerging viral threats, together with antiviral resistance and the limited therapeutic options available for many viral diseases, has intensified the search for faster and more adaptable drug-development strategies. Artificial intelligence (AI) and machine learning are increasingly being integrated across the antiviral discovery pipeline, from target identification and virtual screening to molecular generation, drug repurposing, resistance prediction, and pharmacological optimization. Advances in structural modelling, graph-based learning, generative approaches, and large-scale biological datasets are expanding the chemical and therapeutic space that can be explored computationally.However, improved predictive performance does not necessarily translate into antiviral activity or therapeutic potential. Candidate prioritization remains strongly influenced by data quality, biological context, target flexibility, viral genetic diversity, assay heterogeneity, model uncertainty, and the availability of appropriate external and experimental validation. Consequently, one of the central challenges for AI-enabled antiviral discovery is no longer simply generating predictions, but determining which predictions provide sufficiently robust evidence to guide subsequent experimental and translational decisions.This Research Topic aims to examine how AI-driven approaches can improve the discovery, repurposing, and optimization of antiviral therapeutics while strengthening the connection between computational prediction and biological validation. Particular emphasis will be placed on reproducible and interpretable workflows, prospective evaluation, experimentally validated predictions, resistance-aware modeling, and strategies that integrate computational evidence with pharmacological and virological data. By bringing together computational and experimental perspectives, the collection seeks to identify where AI adds genuine value across the antiviral development pipeline and which methodological advances are required to accelerate progression from in silico prioritization toward experimentally supported therapeutic candidates.We welcome contributions addressing themes including:• AI- and machine-learning-guided antiviral drug discovery• Generative and graph-based models for antiviral compound design• Virtual screening against viral and host therapeutic targets• Computational repurposing of approved and investigational drugs• Structure-based modeling and protein–ligand interaction prediction• AI approaches for antiviral resistance and escape prediction• Prediction and optimization of antiviral combination therapies• Toxicity, pharmacokinetic, and developability prediction• Integration of multi-omics and structural data for antiviral discovery• Prospective and experimental validation of computationally prioritized candidates• Applications to emerging, neglected, and drug-resistant virusesWe welcome Original Research, Methods, and Review articles that provide conceptual, methodological, computational, experimental, or translational advances relevant to AI-enabled antiviral discovery and repurposing. Topic editors: Luis Adrian De Jesús-González, Juan F Osuna-Ramos, VANESSA ALCALÁ-RAMÍREZ Source: https://www.frontiersin.org/research-topics/85723/ai-driven-antiviral-drug-discovery-and-repurposing-from-prediction-to-experimental-validation