saltswap
simulates dynamic anion-cation (salt) explicit solvent environments in OpenMM.
saltswap
uses the semi-grand canonical ensemble to couple a simulation box to a reservoir at a given macroscopic salt concentration.
The applied chemical potential in saltswap
reflects the salt concentration of the reservoir.
@article{saltswap,
author = {Gregory A. Ross, Ari\"{e}n S. Rustenburg, Patrick B. Grinaway, Josh Fass, and John D. Chodera.},
title = {Biomolecular simulations under realistic macroscopic salt conditions},
journal = {bioRxiv},
doi = {10.1101/226001}
year = {2017},
}
To install:
python setup.py install
Dependencies are currently listed in saltswap/devtools/conda-recipe/meta.yaml
.
We'll first create a box water in which to try the insertion and deletion of salt:
from simtk import openmm, unit
from simtk.openmm import app
from openmmtools.testsystems import WaterBox
wbox = WaterBox(box_edge=25.0 * unit.angstrom, nonbondedMethod=app.PME,
cutoff=10 * unit.angstrom, ewaldErrorTolerance=1E-4)
Initializing the sampler that will perform Markov chain Monte Carlo by mixing molecular dynamics moves that transform pairs of water molecules to anions and cations:
from saltswap.swapper import MCMCSampler
sampler = MCMCSampler(wbox.system, wbox.topology, wbox.positions, delta_chem=800)
The chemical potential is specified by delta_chem
and is in kJ/mol.
The chemical potential corresponding to each macroscopic salt concentration must be calibrated for each water and ion pair forcefield, as well as the specific nonbonded treatment.
By default, saltswap
makes instantaneous swaps between water and salt.
To achieve usable acceptance rates, we use nonequilibrium candidate Monte Carlo (NCMC).
To run a simulation for 1000 iterations of MD and salt-swap moves, we use the move
method:
sampler.move(1000)
More detailed examples can be found in examples/
.
- Gregory Ross: [email protected]
- Bas Rustenburg: [email protected]
- John D. Chodera: [email protected]