Evolution of Computational Chemistry Applied to Virtual Drug Screening: A Mini-Review

Abstract

Computational chemistry has become a strategic component of drug design by enabling the evaluation of large chemical libraries, the investigation of molecular interactions, and the prioritization of candidates before experimental testing. This mini-review examines the evolution of computational chemistry applied to virtual screening, highlighting its main approaches, the physicochemical filters used for compound selection, and the incorporation of artificial intelligence-based methods. A narrative review of the literature was conducted using Scopus, Web of Science, PubMed, SciELO, and ScienceDirect, supplemented by Google Scholar. Publications from 2015 to 2026 were prioritized, along with seminal studies on the Lipinski, Ghose, Egan, Muegge, and Veber filters. The literature indicates that structure-based and ligand-based drug design provide complementary strategies for the identification and optimization of bioactive molecules. Methods such as molecular docking, molecular dynamics, QSAR, pharmacophore modeling, and ADMET prediction improve selection efficiency, whereas descriptors such as molecular weight, lipophilicity, solubility, polar surface area, ionization, flexibility, and the fraction of sp³ carbons (Fsp³) aid in assessing pharmacokinetic potential. Machine learning and deep learning have expanded the ability to explore vast chemical spaces and predict molecular properties. However, the reliability of computational results depends on the quality of the structures, data, and models employed. In conclusion, integrating computational tools, physicochemical criteria, and experimental validation can reduce costs, support decision-making, and increase the efficiency of therapeutic candidate discovery.