A protein–protein docking run can produce a long list of poses—and take longer than expected—when the search explores orientations that biology has already made unlikely. But narrowing the search too aggressively can hide a plausible binding mode. The practical challenge is to use what you know about the binding site to focus the calculation without treating an uncertain guess as fact.
Hex, available through its SAMSON Extension, lets you restrict the rotational search with range angles. This is most useful when you have a reasonable idea of where the proteins interact. If the interface is unknown, a broad search may be more appropriate; an angle of 180° corresponds to no angular constraint.
Start with structures you can trust
Before docking, prepare both proteins. Remove alternate-location atoms and, as appropriate, waters, monatomic ions, and other ligands. Add hydrogens if they are absent, but do not change hydrogens that already reflect a chosen protonation state. If residues or heavy atoms are missing, the PDBFixer Extension can help repair the structure. Preparation matters because docking scores and steric checks depend on the atoms in the input models.
In SAMSON, open Hex, assign one protein as the receptor and the other as the ligand, and save the ligand’s initial conformation. That saved state gives you a reference if you later rerun docking with different settings.
Use a known site to focus the search
Hex samples orientations in spherical coordinates. In the Advanced parameters, choose Range angles as the sampling method. If you know the approximate binding site, first move and orient the ligand near that site. Then restrict the receptor and ligand angle ranges. The tutorial uses 45° for each as an example of a loose constraint around a starting orientation.
These angles define spherical cones around the intermolecular axis connecting the protein origins. Smaller ranges mean fewer orientations are considered, which can reduce search time. The trade-off is important: if your starting placement or interface hypothesis is wrong, a narrow range can exclude relevant poses. Use constraints because you have a reason to expect an interface in that region—not simply because a smaller search is faster. You can also limit the ligand’s twist rotation about the intermolecular axis.
Balance speed and resolution
Hex can use a fast, lower-resolution Steric scan before the higher-resolution Final search. The first stage discards many orientations, allowing the more detailed stage to focus on a smaller set. The documentation notes that a Steric scan order of 16–18 is often sufficient, and describes combinations such as 16 for the scan and 25 for the final search as practical starting points. A higher Final search order takes more time; the suggested value is generally 25 for more flexible or model-built structures and 30 for high-resolution crystal structures expected to change little on binding.
These are starting points, not guarantees. Use the Hex interface tooltips and manual for parameter details, and consider the structure quality and flexibility when choosing settings. If you enable post-processing, a bumps counter can help identify poses with steric clashes. Molecular-mechanics refinement is described as alpha-stage functionality in Hex, and its rigid-body approach may be unsuitable when substantial conformational changes are expected.
Inspect poses before drawing conclusions
After docking, review the clustered results rather than treating a top-ranked pose as a confirmed complex. Selecting a row restores that ligand conformation; you can also create conformations in the document and compare them visually. Interaction analysis, distance measurements, hydrogen-bond visualization, and checks for clashes can help you assess whether a pose is plausible. These checks support interpretation, but they do not establish that a predicted interaction occurs experimentally.
The useful habit is to match the search breadth to the evidence you have: keep it broad when the interface is uncertain, and constrain it carefully when you have a justified starting hypothesis. For the complete workflow and parameter details, see the Protein docking with Hex tutorial.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.