Dresden 2026 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
BP: Fachverband Biologische Physik
BP 2: Computational Biophysics I
BP 2.10: Vortrag
Montag, 9. März 2026, 12:00–12:15, BAR/0106
Extending quantum-mechanical benchmark accuracy to biological ligand-pocket interactions — •Mirela Puleva1, Leonardo Medrano Sandonas2, Balázs D. Lorincz3, Jorge Charry4, David M. Rogers5, Péter R. Nagy3, and Alexandre Tkatchenko1 — 1University of Luxembourg, Luxembourg — 2TUD Dresden University of Technology, Germany — 3Budapest University of Technology and Economics, Hungary — 4Luxembourg Researchers Hub asbl, Luxembourg — 5Oak Ridge National Laboratory, USA
Predicting the binding affinity of ligands to protein pockets is key in the drug design pipeline, yet accurate capture of interactions in the flex- ible ligand-pocket motifs requires robust quantum-mechanical (QM) benchmarks, which are scarce. Disagreement between "gold standard" Coupled Cluster (CC) and Quantum Monte Carlo (QMC) methods further challenges large non-covalent benchmarks. We introduce the QUantum Interacting Dimer (QUID) benchmark framework modeling diverse ligand-pocket motifs. CC and QMC agree within 0.5 kcal/mol for QUID, which spans key non-covalent binding motifs and energetic contributions from symmetry-adapted perturbation theory. Bench- mark results shows several dispersion-inclusive density functional ap- proximations predict energies accurately but differ in atomic van der Waals forces, while semiempirical and empirical methods need im- provements for non-covalent interactions (NCIs) in out-of-equilibrium geometries. With a wide span of NCIs, highly accurate interaction en- ergies, and further molecular properties, QUID goes beyond the "gold benchmark" QM benchmark of ligand-protein systems.
Keywords: Protein-ligand interactions; Molecular dimers; Benchmark dataset; Binding energy