Pulling one part of a molecular system away from another can look straightforward: choose two groups, set a direction, and apply a moving restraint. But in a periodic simulation, the box is part of the experiment. If it is too short along the pulling direction, the molecule may interact with its periodic image before it reaches the intended displacement.
This workflow uses the 2BEG system as an example: chain A is pulled away from chain B along the z-axis, while chain B is restrained. The goal is a 5 nm displacement over a 0.5 ns production run. The central planning question is therefore not only “How far do I want to pull?” but also “Does the periodic box have enough room for that motion?”
Plan the box around the motion
For the minimum image convention, the pull distance must be less than half the box length along the pulling direction, taking periodicity into account. For a 5 nm pull in z, the tutorial sets the z length to 12 nm: twice the pull distance gives 10 nm, with an additional 1 nm on each side to provide extra room if parts of the pulled chain extend beyond the nominal displacement.
This is a useful general planning habit: decide the pull direction and target distance before choosing the box. Then check the box length against the half-box limit and leave appropriate space for the system’s shape and movement. If the direction is not aligned with a box axis, orient the system first; SAMSON provides move editors and a Move selection > Align… command for alignment.
In the example, the box is orthorhombic, with lengths 6.5 nm x 4.5 nm x 12 nm. The system is not centered automatically: Center in box is deactivated and the center is set to 0.5 x 0.5 x 0.2. This positioning helps place the system appropriately within the elongated box for the planned z-direction pull. Check the box visualization in the viewport before running the simulation.
Define the two groups deliberately
COM pulling acts on groups, so confirm that the groups correspond to the intended parts of the system. In this example, custom index groups named ChainA and ChainB are created from selections in SAMSON. The group names make the later pull settings easier to inspect, and checking each group selection helps catch mistakes before a long run.
For the production simulation, chain B is position-restrained so it stays in place while chain A is pulled. In the COM pulling parameters, the groups are set to chain A and chain B, the geometry is distance, and the pull direction is N N Y—the z-direction in this setup. The pull is configured with init = 0 nm, a rate of 0.01 nm/ps, and a force constant of 1000 kJ mol^-1 nm^-2. At that rate, a 5 nm displacement takes 500 ps, or 0.5 ns.
Use the plots as a sanity check
After preparation, energy minimization, and NPT equilibration, run the production MD with the pulling settings. The tutorial uses a 2 fs timestep and a 0.5 ns production duration. Once the simulation finishes, inspect the trajectory and the generated plots for pull coordinates and force. The coordinate plot lets you check whether displacement progresses along the intended pull coordinate; the force plot shows how the force changes during the motion. In the example, force builds as the pull overcomes restoring interactions between the chains.
These plots are not a substitute for checking the setup, but they help connect the input parameters to the resulting motion. If the trajectory or pull coordinate does not match the intended direction and displacement, revisit the group definitions, restraints, and box geometry before using the output downstream.
The pulling trajectory can be used as a starting point for umbrella sampling and subsequent potential of mean force analysis. For the complete walkthrough and screenshots, see the GROMACS Wizard COM Pulling tutorial.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON at samson-connect.net.