Pulling one part of a molecular system away from another sounds straightforward—until periodic boundaries enter the picture. In a periodic simulation box, a molecule that travels too far can interact with its own periodic image. A box that is too short in the pulling direction can therefore undermine the setup before the simulation even begins.
This workflow focuses on two practical challenges: choosing a box that can accommodate the intended displacement, and defining the pulling and reference groups clearly. The example uses the 2BEG system, pulling chain A away from chain B along the z-axis.
Plan the box around the movement
For the example, the target displacement is 5 nm. The pull distance must remain less than half the box length along the pulling direction, taking periodicity into account. That means the z dimension needs to be more than 10 nm; the tutorial uses 12 nm, leaving additional room beyond the minimum.
The example box is orthorhombic, with lengths of 6.5 nm × 4.5 nm × 12 nm. Its center is set to 0.5 × 0.5 × 0.2 of the box dimensions rather than centering the system automatically. The point is not to copy these dimensions for every system: choose them in light of the direction and distance you intend to pull, and check that the geometry leaves adequate space under the minimum image convention.
If the intended pull direction is not aligned with a box axis, orient the system before preparing it. In SAMSON, the tutorial suggests using move editors or the structure context menu’s Move selection > Align… command.
Make the groups explicit
GROMACS needs index groups for the two parts involved in the pull. In this example, those groups are chain A and chain B. If they are not among the default index groups, create custom groups in the GROMACS Wizard’s Edit index groups window.
Load the system into SAMSON, select chain A in the Document view, and use Generate based on current selection in document to create its selection string. Name the group ChainA and add it. Repeat for chain B, naming that group ChainB. Before running, select each group in the index-group list and use Select in document based on selection string to check that it identifies the intended chain. This quick check helps catch a group-definition mistake before it becomes a simulation problem.
Set up the pull and restraint as separate choices
The example restrains chain B during production and pulls chain A relative to it. In the production parameters, activate the position restraint for chain B. Then add a COM-pulling coordinate with type umbrella, geometry distance, group 1 ChainA, and group 2 ChainB. Set start to yes, choose N N Y for the distance direction, set the initial distance to 0 nm, and use a force constant of 1000 kJ mol-1 nm-2.
For a 5 nm displacement over 0.5 ns, the rate is 0.01 nm/ps. The tutorial’s production run uses a 2 fs timestep. Keeping the target distance, duration, and rate consistent is a useful setup check: 0.5 ns is 500 ps, and 500 ps at 0.01 nm/ps corresponds to 5 nm.
Check the outcome, not just the input
After the production simulation, load the resulting trajectory in SAMSON to inspect the movement. The generated pull-coordinate and pull-force plots help show how displacement and force change over time. In the example, the force builds as the system responds to the separation; the plots are useful for checking the behavior of the pull, rather than relying only on the values entered in the parameter window.
This workflow produces pulling output that can be used in a subsequent umbrella-sampling workflow. It is a practical bridge between setting up a controlled displacement and preparing for further free-energy analysis—not a substitute for checking whether the resulting trajectory and sampling are suitable for the scientific question.
For the complete walkthrough, see the GROMACS Wizard COM Pulling documentation.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.