Table of Contents
Flow Charts
The Global MD Algorithm
Main phases of a molecular dynamics simulation
The flow of data
MD-parameters
Time Step
The maximum time step with which MD simulations can be expected to be stable depends on the fastest oscillations of the model.
| conditions | typical maximum time step |
|---|---|
| simulation without bond- or angle-constraints | 0.5 fs - 1 fs |
| covalent bonds constrained | 2 fs |
| using virtual interaction sites | ~ 5 fs |
| using a coarse-grained forcefield (e.g. MARTINI) | 20 - 40 fs |
Section “Choosing an appropriate timestep” in Braun-2019
Cut-Offs
The cut-off scheme should match what was used for parameterization of the forcefield. It has been shown that using PME for determining long-range electrostatics in many cases improves the results, even if the forcefield has not been parameterized with PME. (FIXME: citations needed)
FIXME: citations needed
Force Fields
Different force fields have been parameterized with different philosophies and tuned to reproduce different physical properties or with emphasis on different kinds of molecules.
Generally speaking, different forcefield cannot be combined unless they are members of the same family and therefore the secondary forcefield has been specifically parameterized to be compatible with the parent-forcefield.
A Selection of Common Forcefield Families
- AMBER (Assisted Model Building with Energy Refinement)
- GAFF (General AMBER force field) for small molecules
- GLYCAM for carbohydrates
- LIPID21 for lipids
- CHARMM (Chemistry at Harvard Macromolecular Mechanics)
- CGenFF (CHARMM General Force Field) for small molecules
- GROMOS (GROningen MOlecular Simulation)
- OPLS (Optimized Potentials for Liquid Simulations)
- OPLS-UA (OPLS United Atoms forcefield)
- OPLS-AA (OPLS All Atoms forcefield)
- MARTINI - a general purpose Coarse-Grained Force Field