Conformational Ensembles from Experimental Data and Computer Simulations

Conformational Ensembles from Experimental Data and Computer Simulations

Poster Abstracts

31-POS Board 31 Non-Ewald Method for Accurately and Efficiently Calculating Electrostatic Interactions in Molecular Simulations Ikuo Fukuda 1 , Narutoshi Kamiya 2 , Kota Kasahara 3 , Han Wang 4 , Shun Sakuraba 5 , Haruki Nakamura 1 . 1 Osaka University, Suita, Osaka, Japan, 2 University of Hyogo, Kobe, Japan, 3 Ritsumeikan University, Kusatsu, Japan, 4 Institute of Applied Physics and Computational Mathematics, Beijing, China, 5 The University of Tokyo, Tokyo, Japan. A larger system and longer time steps are necessary to conduct a realistic molecular simulation, but they are hardly realized in current computational environments. The most time-consuming part of the simulation is the calculation of long-range interactions of particles. In particular, appropriate treatment of the electrostatic interaction is critical, since the simple truncation cannot be used due to the slow decay of the Coulombic function. Thus, there is strong demand to calculate the electrostatic interactions with high accuracy and low computational cost. For this purpose, we have developed the Zero-multipole summation method (ZMM) [1]. In this method, the periodic boundary condition, which can potentially cause artifacts in particular for heterogeneous systems, is not necessary, and the Fourier-part evaluation, which is typically the bottleneck for high performance computation, is not needed, in contrast to conventional Ewald- based methods. Instead, a suitably defined simple pairwise function, which differs from the original Coulombic function, is used with a cutoff scheme. The underling physical idea is simple that certain electrostatic neutrality is attained in a biological or condensed matter system [2]. This idea is realized by a mathematical foundation to generate a new pairwise function. The accuracy and efficiency of the ZMM has been validated in fundamental systems as well as heterogeneous biomolecular systems, including DNA and protein. In the presentation, we will provide the theory and numerical results on the ZMM, and discuss how the treatment of the electrostatic calculations seriously affects simulation results. [1] I. Fukuda, J. Chem. Phys. 139, 174107 (2013); I. Fukuda et al., ibid. 140,194307 (2014); H. Wang et al., ibid. 144, 114503 (2016); K. Kasahara, et al., Biophys. Physicobiol., 13, 209 (2016). [2] I. Fukuda and H. Nakamura, Biophys. Rev. 4, 161 (2012).

64