Next-generation AI and hybrid simulation approaches for small-molecule inhibitor discovery

Artificial intelligence (AI) and machine learning (ML) have rapidly transformed the landscape of computational drug discovery, enabling data-driven strategies that complement and extend traditional computer-aided drug design. While current AI methodologies showcase remarkable predictive power, a persistent challenge lies in integrating these approaches with physically grounded molecular simulations to achieve deeper mechanistic understanding and improved translational potential. Recent years have witnessed the development of hybrid workflows that merge ML models with molecular docking, molecular dynamics, free-energy calculations, and advanced sampling techniques, yet the field continues to grapple with key questions regarding robustness, interpretability, data quality, and the capacity for real-world impact. As therapeutic discovery demands not just predictive accuracy but also mechanistic insight, there is a pressing need for next-generation, transparent, and translatable AI approaches.This Research Topic aims to catalyze the evolution of AI-driven strategies for the rational discovery, optimization, and repurposing of small-molecule inhibitors. By focusing on the convergence of machine learning and physically informed molecular simulations, the topic encourages contributions that advance our understanding of how digital intelligence can be deployed to both accelerate and deepen insight into the drug discovery process. Core objectives include the development and validation of hybrid computational workflows, the demonstration of model interpretability and robustness across therapeutic areas, and the critical assessment of data integration and bias mitigation strategies. This initiative also seeks to engage with the broader scientific and translational community, illuminating pathways by which AI can yield not only potent drug candidates but also mechanistic clarity and clinical relevance.While this Research Topic concentrates on hybrid AI–physics approaches applied to small-molecule inhibitor development, it embraces a range of innovations spanning algorithmic, methodological, and translational dimensions. Contributions may cover diverse therapeutic contexts — such as oncology, infectious diseases, and neurodegenerative disorders — as well as address foundational challenges in model transparency and applicability. To foster a comprehensive discussion, we welcome original research articles and comprehensive reviews addressing, but not limited to, the following themes:- Hybrid AI–molecular simulation approaches for SMI discovery- Explainable and interpretable ML models in drug design- Deep learning architectures for structure–activity relationship modeling- AI-assisted molecular dynamics and free-energy prediction- Generative AI models for de novo small-molecule design- AI-driven drug repurposing and polypharmacology- Integration of multi-source chemical and biological data for predictive modelling Topic editors: Jose I Bueso-Bordils, Gabriele La Monica, Gerardo M. Anton-Fos Source: https://www.frontiersin.org/research-topics/79681/next-generation-ai-and-hybrid-simulation-approaches-for-small-molecule-inhibitor-discovery