A molecular structure can contain thousands of atoms, but many modeling questions are local: Which receptor residues surround a ligand? Are nearby waters part of the interaction network? Is a contact analysis being distracted by hydrogens or residues that are not relevant? The challenge is not just finding atoms near the ligand—it is expressing the question precisely enough that the selection matches the analysis you intend to do.
SAMSON’s Node Specification Language (NSL) lets you build these selections by combining node types, properties, and spatial relationships. A useful way to approach a complicated selection is to assemble it in layers: first identify the chemical group, then choose the structural level (atoms or residues), then add a distance condition, and finally exclude what you do not want.
Start with the interaction shell
For a first look at a binding pocket, select receptor atoms within 5 Å of the ligand:
(n.t a in n.c rec) w 5A of n.c lig
This query makes the target and reference explicit: the selected nodes are atoms (n.t a) in the receptor (n.c rec), and they must be within 5 Å (w 5A of) of the ligand (n.c lig). Naming both sides is helpful when a structure also contains waters, ions, or other molecules. It makes the intended contact partner clear instead of relying on a broad distance search across the whole model.
If the analysis is about pocket residues rather than individual atoms, change the selected level:
n.t r w 5A of n.c lig
That distinction matters. An atom selection is useful for examining specific contacts; a residue selection is often easier to interpret when reviewing the pocket or considering mutations. To display the atoms belonging to residues in the pocket, the documented pattern is:
n.t a in (n.t r w 5A of n.c lig)
Refine the result for a real question
Many analyses need a more focused set than “everything nearby.” For ligand contact analysis, remove hydrogen atoms from the ligand selection:
(n.t a in n.c lig) and not a.s H
For receptor heavy atoms close enough to flag possible steric clashes, combine non-hydrogen filters on both sides with a tighter distance threshold:
((n.t a in n.c lig) and not a.s H) w 2.2A of ((n.t a in n.c rec) and not a.s H)
This is a screening selection, not a substitute for evaluating the structure and the method used to assess clashes. Its value is that it narrows attention to close heavy-atom pairs, where inspection is likely to be more useful than looking through every ligand–receptor contact.
Other refinements help answer different questions. To inspect polar receptor atoms near the ligand, select oxygen and nitrogen atoms:
(a.s O,N in n.c rec) w 3A of n.c lig
To find nearby waters, use:
n.c wat w 4A of n.c lig
And to focus on water oxygens positioned near both receptor and ligand, combine the two distance conditions:
(a.s O in n.c wat) and (n.t a w 3A of n.c rec) and (n.t a w 3A of n.c lig)
These are useful starting points for inspection, not automatic judgments about whether a water is displaceable or whether a contact is favorable. The selection identifies structural candidates; interpretation still depends on the model and the scientific question.
Make distance filters easier to reason about
When composing a query, read it from the inside out. In (n.t r w 5A of n.c lig) and (r.pI < 6.0), the first condition identifies residues near the ligand, and the second narrows those residues by pI. Parentheses make the intended grouping visible, especially when several conditions are combined.
Likewise, a distance range can describe a shell rather than everything within one cutoff. For example, pocket-rim residues within 12 Å but beyond 7 Å are expressed as:
(n.t r w 12A of n.c lig) and (n.t r b 7A of n.c lig)
This can help separate the immediate binding site from a surrounding region when planning a closer inspection. As you build selections, test the simpler parts first—such as receptor, ligand, or nearby residues—then add atom types, residue properties, and exclusions. That makes it easier to spot whether an unexpected result comes from the target group, the distance condition, or the way filters are combined.
For more examples and the full set of documented patterns, see the NSL examples reference.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.