(360ai) Simulation of Lipid Membranes Using Coarse-Grained Model and Reverse-Mapping

Amphiphilic molecules such as lipids form a wide variety of membranes, including micelles, vesicles, lamellae, or more complex forms. The shape of the membrane depends on the temperature, concentration, and molecular structure of the constituent molecules. [1] Lipid membranes are studied using atomistic molecular dynamics (MD), coarse-grained molecular dynamics, dissipative particle dynamics, and simulations based on continuum models.

Atomistic molecular dynamics methods can provide detailed analytical methods at the atomic level. There are several tools to construct the initial structure of the membrane, such as CharmmGUI [2], Packmol [3],memgen [4], etc. However, often the problem is to know the equilibrium shape of the membrane under a given temperature, pressure and concentration. Moreover, the computational cost for membrane equilibration is very high. CG models can reach equilibrium with less computational effort than atomic models. Using accurate CG potential [5], the membrane structure can be obtained by relaxation simulation without any knowledge of final shape. However, the resolution of the CG model is not always sufficient. Therefore, one approach is to relax the membrane structure using a CG model and reverse-mapping to an atomistic model. This study describes a useful approach for membrane simulations using OCTA system [6] that includes initial particle location, relaxation MD simulation and reverse-mapping from CG to an atomistic model.

1. Arai et al., Phys. Chem. Chem. Phys., 18, 19426 (2016)
2. Jo et al., J. Comput. Chem. 29, 1859-1865 (2008)
3. Martinez et al., J. Comput. Chem., 30, 2157 (2009)
4. Christopher J. Knight and Jochen S. Hub, Bioinformatics, 31, 2897-2899 (2015)
5. Seo and W. Shinoda, SPICA, J. Chem. Theory Comput., 15, 762 (2019)
6. OCTA, http://octa.jp