Two structures can show where a protein starts and ends, but not how it might move between them. That gap is a familiar challenge in molecular modeling: a transition pathway is useful for exploring a conformational change, yet generating one can be difficult when the protein has many atoms and the motion is not obvious.
SAMSON’s Protein Path Finder searches for possible paths between two conformations of the same protein. It uses ART-RRT, combining T-RRT for pathway search with ARAP for motion generation, while constrained minimization helps keep generated motions physically reasonable. The result is a set of candidate paths to inspect—not a claim that a single computed route is the definitive biological mechanism.
Start with a comparable pair of structures
The first practical hurdle is often preparing the inputs. The start and goal conformations need to represent the same system: matching chains, residues, and atoms. If the structures come from separate files, prepare them individually and combine them as different models in one PDB file. For a protein that needs cleanup, SAMSON’s preparation tools can remove alternate locations, ligands, solvent, and ions, and add hydrogens. PDBFixer can help address missing residues or heavy atoms.
For a first attempt, use the tutorial’s sample document. It contains adenylate kinase chain A in two conformations, from the 4AKE and 1AKE structures, already prepared in one document. This lets you learn the workflow before troubleshooting input differences in your own model.
Set up the calculation before searching
Protein Path Finder needs an interaction model and a state updater. The documented example uses Universal Force Field (UFF) and FIRE. The system should already be minimized; the tutorial points to Edit > Minimize for minimization with UFF. When setting up UFF, choose to use the existing bonds. Set FIRE’s step size to 1 fs and its number of steps to 1, as shown in the tutorial.
Next, load the conformations from the active document and assign one as the start and the other as the goal. These choices seed the search, so it is worth checking that the selected structures really are the intended endpoints before running the planner.
Choose which atoms guide the motion
ARAP uses a set of active atoms to control the protein’s motion; the remaining atoms follow according to the method. The tutorial selects two backbone alpha-carbon atoms: CA in GLY 12 and CA in ARG 123. These are examples for the sample system, not universal choices for every protein. Select atoms that make sense for the motion you want to explore, then use Add in the app to assign them as active.
The sampling box also matters: it defines the region in which active atoms are sampled. Its size and position bias their motion and therefore influence the paths found. The tutorial uses a 200 Å cube. Treat that as a starting configuration for this example, not a recommended setting for every system. Check the green box visualization and adjust the region to suit your model and intended motion.
Make runs interpretable
Search parameters determine how the planner explores. In the tutorial, two runs are requested, with 20 ARAP-modeling iterations and 20 constrained minimization iterations for each newly generated state. The T-RRT settings include an initial temperature of 0.001 K, a temperature factor of 2, and one failure before increasing the temperature; the RRT extension step size is 1 Å. These values reproduce the illustrated setup, but they are not a universal recipe.
If you want to repeat a calculation, enable Use seed and record the seed value. The planner increments the seed after each run; without this option it uses a random seed. You can also choose whether to align accepted states to the start conformation. Alignment can make the search faster, but the documentation notes that it may produce artifacts in rare cases, so inspect the resulting motion rather than relying on the setting alone.
Read the results with care
Paths appear in the Results tab as they are found. The table reports the number of states, minimum and maximum energies, and other path information. Select a path to view its energy curve, then move the slider to inspect individual conformations in the viewport. Multiple paths can be selected to compare their curves.
Energy values can help compare the computed conformations within this setup, but they should not be mistaken for a complete description of biological likelihood or experimental kinetics. Review the actual structures and motion as well as the table. If a path is useful, export it as a trajectory or export conformations at a chosen interval. Candidate paths can also be refined with the P-NEB app; see the linked tutorial for that workflow.
This workflow is most useful when you treat path finding as exploration: prepare equivalent endpoints, make deliberate choices about active atoms and sampling space, and inspect the paths before drawing conclusions. For the full instructions and parameter screenshots, see the Protein Path Finder documentation.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.