A center-of-mass pulling simulation can be configured correctly and still produce misleading results if the periodic box is too small along the pulling direction. Under periodic boundary conditions, a molecule that moves across one side of the box can appear again on the other. The practical challenge is to allow the intended displacement without letting periodic images interfere.
This workflow uses the 2BEG system as an example: chain A is pulled away from chain B along the z-axis. The key preparation lesson is simple: plan the box around the motion you intend to impose, not just around the system’s starting shape.
Size the box for the displacement
For this example, the target pull distance is 5 nm. The distance pulled must remain less than half the box length in the pulling direction, taking periodicity into account. That means the z dimension needs to be at least 10 nm. The tutorial adds extra clearance—1 nm on each side—giving a z box length of 12 nm. The resulting box lengths are 6.5 nm × 4.5 nm × 12 nm.
This is a useful planning habit for other systems too: identify the pull axis, estimate the full displacement, and check the box length against the minimum-image constraint before running the simulation. If the pulling direction is not aligned with a box axis, orient the system appropriately in SAMSON before setting up the box.
In GROMACS Wizard, choose an orthorhombic unit cell and compute the fitted box first. Then set the box lengths manually. For the example, deactivate Center in box and place the system’s center at 0.5 × 0.5 × 0.2 of the box dimensions. This positions the system within the larger z dimension rather than simply centering it there.
Define the groups before configuring the pull
The pull coordinate needs clearly identified groups. In the example, create custom index groups named ChainA and ChainB from the corresponding chain selections in SAMSON. Before proceeding, use the option to select nodes in the document from each group’s selection string. This quick check helps confirm that each index group represents the intended chain.
During production setup, apply position restraints to chain B so it stays in place while chain A is pulled. Then add a COM pulling coordinate with these settings:
- Type: umbrella; geometry: distance.
- Group 1: chain A; group 2: chain B.
- Start: yes; direction:
N N Y(the z-axis). - Initial distance:
0 nm; rate:0.01 nm/ps. - Force constant:
1000 kJ mol^-1 nm^-2.
At that rate, a 0.5 ns production run corresponds to a 5 nm displacement. The rate, duration, and box size should therefore be considered together: changing the intended displacement may mean revisiting the box dimensions as well.
Check the outcome, not just the settings
After the simulation, inspect the trajectory to see how the chains move. The production run also generates plots of the pull force and coordinates. Together, these help you review the displacement over time and see how the force changes as the pull proceeds. The output can then be used as a starting point for umbrella sampling.
The full workflow also includes preparation, energy minimization, and NPT equilibration before production. For the 2BEG example, the tutorial uses AMBER03, TIP3P water, neutralizing ions, and 100 mM NaCl; during NPT equilibration, it sets the temperature to 310 K and applies position restraints to chains. These steps help establish the system before the pulling parameters are used.
For the complete instructions and screenshots, see the GROMACS Wizard COM Pulling tutorial.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.