Hybrid Monte Carlo-Molecular Dynamics (MCMD)

GOMC Requirements

GOMC currently requires that Box length / 2 excede the radius of gyration of all molecules in the system.

Grand-Canonical Molecular Dynamics (GCMD) or Gibbs Ensemble with Molecular Dynamics changes the number of molecules in each box. This will alter the ordering of the molecules, posing a challenge when the user tries to concatenate the trajectories or follow one atom through a trajectory.

The GOMC checkpoint file will reload the molecules in the original order to ensure the GOMC trajectories (PDB/DCD) have a consistent ordering for analysis. Atoms that are not currently in a specific box are given the coordinate (0.00, 0.00, 0.00). The occupancy value corresponds to the box a molecule is currently in (e.g. 0.00 for box 0; 1.00 for box 1).

The restart binary coordinates, velocities, and box dimensions (xsc) from NAMD need to be loaded along with the checkpoint file, restart PDB, and restart PSF from the previous GOMC cycle.

The python script from the py-MCMD git repository, combine_data_NAMD_GOMC.py, requires the GOMC step reset to 0 every cycle

InitStep          0

NAMD Requirements

GOMC outputs all the files needed to continue a simulation box in Molecule Dynamics (pdb, psf, xsc, coor, vel, xsc). These files should all be used.

There are certain flexibilities that NAMD allows for that GOMC doesn’t support. To ensure the two systems are compatible the following settings in the NAMD configuration file are required:

Rigid bonds, since GOMC doesn’t support bond length sampling.

rigidBonds          all

Fixed volume, since GOMC maintains the origin of the box at [box length/2, box length/2, box length/2]

# Constant Pressure Control (variable volume)
langevinPiston        off

useGroupPressure      yes

useFlexibleCell       no

useConstantArea       no

Box origin must be centered at [box length/2, box length/2, box length/2]

cellOrigin        x_box length/2      y_box length/2          z_box length/2

Dynamic Subvolumes for Dual Control Volume Molecular Dynamics

To define a subvolume in the simulation, use the subvolume keywords to choose an subvolume id, center, either the geometric center of a list of atoms (dynamic) or absolute cartesian coordinate (static), and dimensions. The residues that can be inserted/deleted in the subvolume, custom chemical potential, and periodicity of the subvolume may also be specified. Fugacity can be replaced for chemical potential. A chemical gradient can be established in the simulation by defining two or more subvolume with different chemical potentials of a given residue. After the molecule is inserted/deleted within one subvolume, it can diffuse to the low concentration subvolume where it is deleted maintaining the concentration gradient via two (2) difference chemical potentials.

To define two control volumes forming a gradient from the left to the right of the box

SubVolumeBox                0       0

SubVolumeDim                0       left_one_fifth y_dim_box_0 z_dim_box_0

SubVolumeResidueKind        0       DIOX

SubVolumeRigidSwap          0       true

SubVolumeCenter             0       left_center y_origin_box z_origin_box

SubVolumePBC                0       XYZ

SubVolumeChemPot            0       DIOX    -2000


SubVolumeBox                1       0

SubVolumeDim                1       right_one_fifth  y_dim_box_0 z_dim_box_0

SubVolumeResidueKind        1       DIOX

SubVolumeRigidSwap          1       true

SubVolumeCenter             1       right_center y_origin_box z_origin_box

SubVolumePBC                1       XYZ

SubVolumeChemPot            1       DIOX    -4000

Run a Hybrid Monte Carlo-Molecular Dynamics Sim

GOMC and NAMD produce compatible input/output files, which allow the system to alternate between Monte Carlo and Molecular Dynamics. The py-MCMD script automates the directory generation, running of GOMC and NAMD, and concatenation of the short alternating runs. Simulating the Grand Canonical ensemble in GOMC with only molecule transfers, allows the MD simulations to continue where they left off, with a varying number of molecules.

Refer to the section on Hybrid Monte Carlo-Molecular Dynamics in the manual and attached links.

Link to documentation: https://py-mcmd.readthedocs.io/en/latest/

Link to Github Repository: https://github.com/GOMC-WSU/py-MCMD

$ git clone https://github.com/GOMC-WSU/py-MCMD.git
$ cd py-MCMD
### Run hybrid simulation
$ python run_NAMD_GOMC.py -f user_input_NAMD_GOMC.json
### Combine alternating GOMC/NAMD cycles into two single GOMC and NAMD data and trajectories.
$ python combine_data_NAMD_GOMC.py -f user_input_combine_data_NAMD_GOMC.json