A ligand can fit neatly into a binding site and still leave a frustrating question: which route might it take out? A static structure shows where the ligand is bound, but not how it could move through a protein while the surrounding atoms adjust. Manually dragging a ligand is useful for visualization, yet it is difficult to explore many plausible routes systematically.
SAMSON’s Ligand Path Finder addresses this problem by searching for candidate unbinding pathways with ART-RRT. The approach combines T-RRT pathway search with ARAP modeling to generate ligand motion, while constrained minimization adapts the protein as the ligand moves. The result is a set of candidate paths that you can inspect as conformations and energy profiles in the same molecular model.

Start with a model that is ready to move
Before searching, use a minimized protein–ligand system. The tutorial’s sample contains lactose permease and the ligand thiodigalactoside (TDG), with a minimized conformation named bound_minimized. For a system of your own, preparation may include removing alternate locations, water, and ions, adding hydrogens, and fixing structural issues. The quality and orientation of the starting model matter: the search region is a box in Cartesian coordinates, so orient the protein so the box can cover the exits you want to explore.
The workflow uses the Universal Force Field (UFF) as its interaction model and the FIRE state updater. In the example setup, FIRE uses a step size of 1 fs and one step. These choices establish the energy evaluation and updating setup for the search; they do not remove the need to inspect the model and results carefully.
Tell the search what can move
Three selections define the search problem:
- Starting conformation: choose the bound, minimized state from which the search tree begins.
- Ligand atoms: select the complete ligand. The remaining atoms are treated as protein atoms.
- Active and fixed ARAP atoms: active ligand atoms guide motion, while passive ligand atoms follow them. Fixed protein atoms help prevent the protein from drifting along with the ligand.
In the sample, the sulfur atom S1 in TDG is selected as an active atom. A protein backbone CA atom in HIS 205 is used as a fixed atom. These are example choices, not universal prescriptions: which atoms you designate can affect the pathways found. Choose fixed atoms in regions you expect to be relatively static, and check the selections in the app before running.

The box is part of the scientific question
The sampling box defines where the active atoms can be sampled. Its position and dimensions therefore bias the ligand’s motion and the routes the search can discover. The app suggests a box enclosing the ligand and protein atoms, but that suggestion is a starting point—not a substitute for choosing a region that covers the exit directions you want to examine. In the tutorial, the box is set to bias motion toward the periplasmic side.

Balance exploration, repeatability, and time
Search parameters control how much exploration is attempted and how each sampled state is treated. The example uses two runs, 20 ARAP-modeling iterations, 20 constrained-minimization iterations, and a 1 Å RRT extension step. It also stops a run when the ligand center has moved 40 Å from its initial position or when the maximum time is reached.
If you need to reproduce a search, enable Use seed and record the seed value along with the other settings. If the option is unchecked, the planner uses a random seed. A fixed seed supports repeatability under the same setup; it does not make the result independent of choices such as the starting structure, fixed atoms, or sampling box.
Read the paths as candidates, not just animations
Each found path appears in the Results tab. The table reports the number of states, minimum and maximum energies, saddle and barrier values, elapsed search time, and the energies of the first and last conformations. Selecting a path displays its energy curve; moving the slider updates the molecular model to the corresponding conformation. You can select multiple paths to compare their curves.

These outputs help you compare the routes produced by a particular setup. A path is a computational candidate, so examine both its conformations and its energy profile rather than treating the table’s lowest value as a verdict on biological relevance. The search box, atom assignments, force field, and minimization settings all shape what is explored. If a route looks useful, you can export paths or conformations for further analysis; the documentation also describes refining paths with P-NEB.
For the complete step-by-step workflow, parameter figures, and export instructions, see the Ligand Path Finder tutorial.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.
