The effective fragment potential (EFP) method is a computational approach designed to simulate accurate intermolecular interactions and solvent effects. Recently, we developed an extension of the EFP method to simulate non-covalent interactions in proteins, called BioEFP. In this method, we make use of a divide-and-conquer approach to partition a biomolecule into smaller ‘effective fragments’. The parameters corresponding to these effective fragments are obtained using a set of preparatory ab-initio calculations. In the polarizable quantum mechanical (QM)/BioEFP simulations, the primary system of interest, which comprises of the chromophore and relevant residues that directly interact with the chromophore, are described using QM methods, while the surrounding environment is described using BioEFP. In recent benchmark studies, we compared the accuracy of the QM/BioEFP method for simulating the electron excitation, electron affinity and vertical ionization energies, with the full QM description of the system. In this work, we compare the properties of proteins obtained with QM/BioEFP and the Fragment Molecular Orbital (FMO) methods. Our results indicate that QM/BioEFP accurately reproduces the properties of large molecules at a fraction of the cost of the QM methods. For instance, in the α-helix form of the alanine polypeptide (ALA)10, the dipole moments of the individual fragments and that of the whole polypeptide obtained using BioEFP and FMO show a very good correspondence. As expected, polarization plays a key role in simulating the properties of biomolecules. Tuning the short-range EFP-EFP polarization damping holds the key towards reproducing the properties obtained using the QM methods. Additionally, we use big data analytics to understand the dependence of EFP parameters with respect to change in conformational geometries, in order to explore the transferability of EFP fragments. Once implemented, the combination of transferable EFP and BioEFP is expected to provide a robust universal force field, capable of describing intermolecular interactions at both short and long range scales.

Towards development of the ab-initio-based force field for biological systems

253rd National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.