Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

CPP: Fachverband Chemische Physik und Polymerphysik

CPP 9: French-German Session: Simulation Methods and Modeling of Soft Matter I

CPP 9.5: Vortrag

Montag, 9. März 2026, 16:15–16:30, ZEU/LICH

Bridging Accuracy and Sampling: Insights into dAMP Solvation from ML-potentials — •Laurie Stevens^1,2, Riccardo Martina², Alberta Ferrarini², and Marialore Sulpizi¹ — ¹Chair of Theoretical Physics of Electrified Liquid-Solid Interfaces, Faculty of Physics and Astronomy, Ruhr-Universität Bochum, Germany — ²Department of Chemical Sciences, Università degli Studi di Padova, Italy

Although nucleotides are essential biomolecular building blocks, key aspects of their structure and interactions, in particular the steps enabling their assembly and polymerization, remain unclear. Here, we investigate the conformational behavior of deoxyadenosine monophosphate (dAMP) in solution using a combined computational-experimental approach. We developed a highly accurate machine-learning potential via active learning that thoroughly samples the free-energy landscape, capturing critical conformational degrees of freedom with near-chemical accuracy. This potential enables nanosecond-scale simulations of water and nucleotides at a level previously inaccessible with hybrid DFT methods. We find that, in solution, dAMP accesses both anti and syn glycosidic conformations. We identify the anti and high anti conformers as the most stable ones, we show that they underlie the intensity variations observed in mid-IR experimental spectra.

Keywords: Machine Learning; Molecular Dynamics; Neural Network Potential; Nucleic Acids; Computational Spectroscopy