Two structures of the same protein can show different conformations—but getting from one to the other is rarely as simple as interpolating between coordinates. A straight-line blend can produce distorted structures, and it does not explain what intermediate motions might connect the endpoints. For molecular modelers exploring a possible conformational transition, the practical challenge is to generate candidate paths that can be inspected, compared, and refined.
SAMSON’s Protein Path Finder provides one way to search for such paths. It uses ART-RRT: T-RRT for pathway search and ARAP for generating motion, with constrained minimization applied during the search. The result is a set of candidate transitions to explore—not a claim that the protein follows that exact route in nature.
Start with matched structures
The two endpoint conformations need to describe the same molecular structure: matching chains, residues, and atoms. If your structures are in separate files, prepare them individually first, then combine them as separate models in a single PDB file. For a new workflow, the tutorial’s sample document is a useful way to learn the controls before bringing in your own system.
Preparation matters. The documentation recommends removing alternate locations, ligands, solvent, and ions, and adding hydrogens. If residues or heavy atoms are missing, the PDBFixer extension can help repair the protein. The sample used in the tutorial is already prepared and contains two conformations of adenylate kinase: 4AKE as the start and 1AKE as the goal.
Set up an energy model before searching
The system must already be minimized. In the example workflow, the interaction model is Universal Force Field (UFF), and the state updater is FIRE. When prompted during UFF setup, choose to use existing bonds. The example also sets FIRE’s step size to 1 fs and its number of steps to 1.
These choices are part of the tutorial’s setup, not universal settings for every protein or research question. In particular, the energy values shown later are calculated with the selected interaction model. Treat them as values for comparing paths within this setup, not as experimental measurements or direct estimates of biological free-energy barriers.
Choose what controls the motion
Protein Path Finder asks you to choose active ARAP atoms. These atoms guide the motion; the remaining atoms follow according to the ARAP method. In the sample, two backbone alpha-carbon atoms are selected: the CA atoms in GLY 12 and ARG 123. The document includes a group for these atoms, so you can select it and use Add in the app.
This is a useful modeling decision to make deliberately: the selected atoms influence how the motion is steered. After adding them, check the active-atom count in Advanced information, and use the Select controls to verify which atoms were assigned. The sample also visualizes active atoms in green.
Give the search room, then tune it
The sampling box defines the region in which the active atoms can be sampled. Its size and position bias their motion, so it can affect the paths you find. The tutorial uses a cube measuring 200 angstroms on each side. That is a sample setting, not a recommended value for every system; inspect the visualization and choose a region appropriate to your setup.
Search parameters govern the trade-off between exploration and computation. The example runs the search twice, with 20 ARAP-modeling iterations and 20 constrained-minimization iterations per generated state. A seed can make a run reproducible under the same setup; without one, the app uses a random seed. The maximum elapsed time per run can limit how long each attempt continues. During a run, monitor nodes, elapsed time, run number, and paths found in Advanced information.
Inspect candidates instead of judging by one number
Found paths appear in the Results tab. Select a path to inspect its energy curve, then move the slider to view individual conformations in the viewport. You can select multiple paths to compare their curves. The table reports quantities such as the number of states, minimum and maximum energy, saddle, barrier, and search time; these make it easier to organize and compare candidates, but do not replace visual inspection or scientific validation.
When a path is useful, export it as a trajectory in the document, or export selected conformations at a chosen interval. You can also copy the path table or energy values for further analysis. For additional refinement, the documentation points to the P-NEB workflow; pathlines offer another way to visualize molecular movement.
For the complete setup instructions and interface details, read the Protein Path Finder tutorial.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.