A center-of-mass pulling simulation can look straightforward: choose two molecular groups, pick a direction, and apply a pulling coordinate. But there is a less obvious source of trouble: the periodic box. If it is too short along the pulling direction, a molecule may interact with its periodic image before it has traveled the distance you intended. The result can be difficult to interpret, even when the pulling parameters themselves look reasonable.
This workflow uses the 2BEG system to illustrate how to plan the box and set up a simple COM pulling run in GROMACS Wizard. The example pulls chain A away from chain B along the z-axis, while restraining chain B.
Start with the distance and the box
For this example, the target displacement is 5 nm. Under the minimum image convention, the pull distance must be less than half the box length in the pulling direction. That means a 10 nm z-dimension is only the theoretical minimum for a 5 nm pull. The tutorial allows additional room: 1 nm on either side, for a total z-length of 12 nm.
The example box is orthorhombic, with lengths of 6.5 nm x 4.5 nm x 12 nm. Its center is set to 0.5 x 0.5 x 0.2 of the box dimensions rather than centered automatically. The key planning idea is to consider both the group’s intended displacement and the space occupied by the molecule as it moves—not just the initial distance between groups.
If the desired pulling direction is not aligned with a box axis, orient the system before defining the box. SAMSON provides ways to move or align a selection; the important point is that the chosen direction and the box dimension you size for it should agree.
Define the groups you actually want to move
Pulling depends on the groups used to calculate the centers of mass. In the 2BEG example, the required groups are chain A and chain B. If these are not among the default GROMACS index groups, add custom groups in GROMACS Wizard. Load the system into SAMSON, select chain A in the Document view, generate a group from the current selection, name it ChainA, and add it. Repeat for chain B as ChainB.
It is worth checking each group after creating it: select it in the index-group list and use the option to select the corresponding nodes in the document. This simple verification helps catch a group-selection mistake before it becomes a confusing simulation result.
Set up the pull and restraint consistently
After preparation, minimization, and NPT equilibration, configure the production simulation. In this example, chain B receives position restraints so it stays in place while chain A is pulled. In the COM pulling settings, add a pull coordinate with:
- Type:
umbrella; geometry:distance. - Group 1: chain
A; group 2: chainB. - Start:
yes; direction:N N Y(the z-direction). - Initial distance:
0 nm; rate:0.01 nm/ps. - Force constant:
1000 kJ mol^-1 nm^-2.
The rate matches the planned motion: 0.01 nm/ps for 500 ps gives 5 nm, and 500 ps is 0.5 ns. Checking this unit conversion is a useful safeguard when translating a desired distance and simulation duration into a pulling rate.
Check what the run produced
The example production run is 0.5 ns with a 2 fs timestep. Once it finishes, the trajectory can be loaded into SAMSON to visualize the motion. The generated pulling plots show the pull coordinates and force over time. These are useful for checking whether the coordinate changed as expected and for seeing how the force developed during the pull.
The pulling output can also serve as a starting point for umbrella sampling, where configurations along the path are used to create sampling windows for free-energy analysis. Pulling is therefore a setup step, not by itself a complete free-energy calculation.
For the complete workflow and interface details, see the GROMACS Wizard COM Pulling tutorial.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.