Quantum Pharmacophore-Based Virtual Screening Enables Prospective Discovery of Chemotype-Diverse Dengue NS5 Inhibitors.

In this work, we introduce a quantum pharmacophore framework that fundamentally redefines how molecular interactions are represented and screened computationally. Unlike traditional pharmacophore or docking-based approaches that rely on empirical feature definitions, conformational sampling, and Cartesian coordinates, our method derives pharmacophores directly from density functional theory and quantum theory of atoms in molecules. This yields a target-conditioned, topological, and interaction-centric representation of ligand-target complexes, enabling rigorous dimensionality reduction and substantial computational acceleration. To our knowledge, this is the first demonstration of a quantum-derived topological pharmacophore capable of supporting subgraph isomorphism-based virtual screening at chemical library scale. We prospectively applied quantum pharmacophore-based screening to three conserved pockets of the dengue virus NS5 RNA-dependent RNA polymerase, evaluating 43.8 million compounds and identifying five chemically diverse inhibitors validated in biochemical, biophysical, and cellular assays. Notably, three compounds target the highly dynamic NITD-640 pocket with biochemical IC50 values in the low- to mid-micromolar range (∼3 to 123 μM), improved ligand efficiency, and drug-like properties relative to the reference ligand. Together, these results demonstrate that quantum pharmacophores can uncover dissimilar chemotypes at challenging, flexible binding sites that are poorly addressed by conventional screening methods.