A ligand’s route out of a protein binding site is not always obvious. A straight-line displacement may pass through protein atoms, while the surrounding protein can also adjust as the ligand moves. For molecular modelers, the challenge is to explore possible exit routes in a controlled, inspectable way—and to understand how setup choices affect what the search finds.
SAMSON’s Ligand Path Finder provides one workflow for exploring candidate unbinding pathways. It uses ART-RRT, combining T-RRT pathway search with As-Rigid-As-Possible (ARAP) modeling to generate ligand motion. Constrained minimization is also used to adapt the protein during the search. The result is a set of candidate paths to inspect, not a guarantee that a path is the unique or experimentally confirmed route.
Start with a model you can trust
Before searching, make sure the system is prepared and minimized. The tutorial’s example is lactose permease with the ligand thiodigalactoside (TDG). It uses a minimized bound conformation, bound_minimized, as the starting state. For your own structure, preparation may include removing alternate locations, waters, and ions, adding hydrogens, and fixing structural issues as needed. The documentation points to SAMSON’s protein preparation tutorial for those steps.
The workflow also requires the Ligand Path Finder app and the FIRE state updater. In the app’s Settings, choose Universal Force Field (UFF) as the interaction model and FIRE as the state updater. The system should already be minimized; FIRE is then configured with a 1 fs step size and one step for this workflow.
Make the search setup deliberate
Three selections define how the method treats the system:
- Starting conformation: select the bound, minimized conformation. This becomes the search tree’s starting state.
- Ligand atoms: select the ligand and set it in the app. The remaining atoms are treated as protein atoms.
- Active and fixed ARAP atoms: active ligand atoms control the motion, while the other ligand atoms follow them. Fixed protein atoms help prevent the protein from drifting along with the ligand.
These choices are not merely bookkeeping. The documentation notes that fixed-atom selection can influence the resulting pathways, so choose atoms in protein regions expected to be relatively static. The active atoms likewise determine how ligand motion is controlled by the ARAP model. In the tutorial system, one sulfur atom in TDG is used as the active atom, and a protein backbone CA atom is fixed.
Next, define the sampling box. It bounds the region in which active atoms are sampled, so its position and dimensions bias the ligand’s motion and therefore the paths that may be found. Check that the box covers the exit region you want to explore. Also orient the system with the Cartesian box in mind: the tutorial recommends aligning the model so the search domain efficiently encloses plausible exits.
Balance repeatability and exploration
Search settings affect both the search process and how easy it is to compare runs. A specified seed can help reproduce results for the same setup; with the seed option enabled, the value increments after each run. Without it, the planner uses a random seed.
The tutorial’s example uses two runs, 20 ARAP-modeling iterations, and 20 constrained-minimization iterations for each sampled state. It sets the maximum ligand displacement to 40 Å and the RRT extension step size to 1 Å, with a maximum time per run. These are example settings, not universal defaults for every molecule. Treat them as a starting point and document the choices you make so that comparisons between searches remain meaningful.
Read the results as candidates
As paths are found, they appear in the Results tab. The table reports quantities including the number of states, minimum and maximum path energies, the energy difference between maximum and minimum, and the difference between the maximum and first-state energies. Selecting a path displays its energy curve; moving the slider updates the corresponding conformation in the structural model.
Use these views together. A path’s energy summary is useful for screening, while inspecting its conformations lets you see how the ligand moves through the protein. Multiple paths can be selected to compare their energy curves. You can also export path data, trajectories, or conformations for further analysis. If a candidate needs refinement, the documentation describes using the P-NEB app to improve paths.
The practical takeaway is to interpret a computed route in the context of its setup: starting structure, ligand and atom assignments, fixed atoms, sampling-box placement, and search parameters all matter. For the full step-by-step workflow, see the Ligand Path Finder tutorial.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.