Radial Distribution Functions for GC-AdResS

Purpose of Module

One purpose of our project is to promote GC-AdResS as method which provides new insights and is not much more complex and difficult to use. In GC-AdResS simulations we introduce artificial interfaces, from atomistic to hybrid and hybrid to coarse grained. To make sure that we indeed have an open system we have to check several properties, from structural to dynamic properties. Radial distribution functions (RDF’S) are the easiest way to check the structural properties of the simulation. This module is dedicated to describe a straight forward and easy way to generate them for the atomistic regions in the GC-AdResS simulations. In the current implementation in GROMACS we have two geometric setups. One is radial and the other is a slab like structure. We use VMD as the tool sof choice to calculate the RDF’s.

Background Information

The most widespread tool for analysing molecular dynamics simulations is VMD). The program is based on TCL and Tk scripting language. Documentation and tutorials can be found here: VMD Docs

The reference coordinates (center of the AdResS region, as defined in the GROMACS mdp input file), configuration (standard input GROMACS: conf.gro) and trajectory (standard output GROMACS: traj_comp.xtc) are necessary to run AdResS. And they have to be used for two essential analysis parts for AdResS. The structural part, the radial distribution functions of the simulated system. And the second part with the help of the Density Profile tool (repository contains a tutorial on how to use) show the diffusion of the molecules during the simulation. The RDF plugin is described here: https://www.ks.uiuc.edu/Research/vmd/plugins/gofrgui/. All results can be stored as plain ASCII files, thus can be analysed via any plotting program.

CASE 1: RDF

  • Since we apply the routine in a subspace, the normalization factor is wrong. The correcting factor can be obtained by calculating the RDF in the full atomistic case. Then the RDF in the same subspace (as in the AdResS) in the full atomistic case. The quotient can be used to re-normalize the AdResS RDF.
  • Important: for the AdResS RDF the update selection option tick has to be used, and the PBC switched off.
  • good case: the RDF’s are exactly the same
  • bad case: the RDF’s are different somehow

CASE 2: Density Diffusion

  • It is possible to define the different regions in the simulation box. Thus it is possible to look at the region specific density diffusion.
  • To do that: one has to specify several time frames and calculate the average profiles, each one gives a ASCII file which (in the end) can be plotted together.
  • The frame of reference is set by the slider in the main menu. (The update selection option has to be switched off.) The particles in that frame are tagged and then via the different time frames one follows their path.
  • Good case: smooth regular and symmetric diffusion
  • bad case: spikes at the interface and asymmetric diffusion might hint at an artificial interface between the different regions.

Example Collection

We basically work with the atom selection and use the pre-existing Radial Pair Distribution Function tool in VMD. One has to load the configuration file (standard is conf.gro) and the trajectory file (standard: traj_comp.xtc) via either vmd conf.gro traj_compt.xtc or the vmd GUI.

Example case (for the radial systems), this selection was used with the reference point being defined in the GROMACS input file.

x_ref = x-value of the center of the chosen AdResS region
y_ref = y-value of the center of the chosen AdResS region
z_ref = z-value of the center of the chosen AdResS region

radius: radius of the atomistic region
name "insert your choice of atom here" and (((x-x_ref)^2 + (y-y_ref)^2 + (z-z_ref)^2) < radius*radius )

For the slab structures:

at_start: start of the atomistic regions along the x axis
at_end: end of the atomistic regions along the x axis
hy_start: start of the hybrid regions along the x axis
hy_end: end of the hybrid regions along the x axis

name "insert your choice of atom here" and (x>at_start and x<at_end)
name "insert your choice of atom here" and ((x>hy_start and x<at_start) or (x>at_end and x<hy_end)