Evaluation of Kirkwood-Buff integrals via finite size scaling: a large scale molecular dynamics study

3rd International Conference on Mathematical Modeling in Physical Sciences

物理学;数学

Dednam, W.^1,2 ; Botha, A.E.^2

Departamento de Física Aplicada, Universidad de Alicante, San Vicente del Raspeig, Alicante

E-03690, Spain^1

Department of Physics, Science Campus, University of South Africa, Private Bag X6, Florida Park

1710, South Africa^2

关键词: Conformational transitions;    Finite size scaling;    Finite-size scaling method;    Kirkwood-Buff integrals;    Large-scale molecular dynamics;    Molecular distribution;    Radial distribution functions;    Structure-function relationship;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/574/1/012092/pdf
DOI  :  10.1088/1742-6596/574/1/012092

来源: IOP

【 摘 要 】

Solvation of bio-molecules in water is severely affected by the presence of co-solvent within the hydration shell of the solute structure. Furthermore, since solute molecules can range from small molecules, such as methane, to very large protein structures, it is imperative to understand the detailed structure-function relationship on the microscopic level. For example, it is useful know the conformational transitions that occur in protein structures. Although such an understanding can be obtained through large-scale molecular dynamic simulations, it is often the case that such simulations would require excessively large simulation times. In this context, Kirkwood-Buff theory, which connects the microscopic pair-wise molecular distributions to global thermodynamic properties, together with the recently developed technique, called finite size scaling, may provide a better method to reduce system sizes, and hence also the computational times. In this paper, we present molecular dynamics trial simulations of biologically relevant low-concentration solvents, solvated by aqueous co-solvent solutions. In particular we compare two different methods of calculating the relevant Kirkwood-Buff integrals. The first (traditional) method computes running integrals over the radial distribution functions, which must be obtained from large system-size NVT or NpT simulations. The second, newer method, employs finite size scaling to obtain the Kirkwood-Buff integrals directly by counting the particle number fluctuations in small, open sub-volumes embedded within a larger reservoir that can be well approximated by a much smaller simulation cell. In agreement with previous studies, which made a similar comparison for aqueous co-solvent solutions, without the additional solvent, we conclude that the finite size scaling method is also applicable to the present case, since it can produce computationally more efficient results which are equivalent to the more costly radial distribution function method.