Contact mapper (geometric)

Contact mapper (geometric)

What it does

Takes one structure, a target chain, and a binding partner, and reports every residue in the target chain that comes within a distance cutoff of the partner. This is the fast, repeatable version of manually inspecting an interface in ChimeraX.

The partner can be another chain, a bound ligand, a peptide, or a cofactor. The calculation is the same in every case — what changes is which atoms count as "the partner".

How contacts are decided

A residue is reported if any of its heavy atoms is within the cutoff of any heavy atom of the partner. The closest such approach is what's reported as the distance, along with which two atoms it was between.

Hydrogens are excluded deliberately. Most X-ray structures don't resolve them and most computational models do, so including them would make the same cutoff mean different things depending on where the file came from.

The 4 Å default is the conventional threshold for "in contact" — close enough for van der Waals interaction. Widening it to 5 Å catches more peripheral residues; tightening to 3.5 Å gives a more conservative core. There's no universally correct value, which is why it's a setting and is recorded on every run.

Reading the output

One row per contacting residue:

• PDB residue / residue name — numbering as deposited in this structure.
• UniProt accession / position — the same residue in the canonical protein sequence, via SIFTS. This is what lets contacts from two different structures be compared: PDB numbering varies between entries, UniProt numbering doesn't.
• Closest distance and the two atoms involved.
• Warning — flags a residue with no UniProt mapping, which happens for engineered tags, expression artefacts, and non-standard residues.

If the UniProt columns are empty across the board, SIFTS had no mapping for this entry (common for user-uploaded structures with no PDB ID). PDB numbering is still correct; only the cross-structure comparison is unavailable.

Important limitations

This reports proximity, not importance. A residue that touches is not the same as a residue that drives binding. Distinguishing contacts from energetic hotspots needs alanine-scanning-style analysis, which this version doesn't do.

One method, no cross-validation. Different contact-detection methods disagree at the margins. A rigorous interface definition usually takes residues found by more than one method. This tool gives you one method's answer — treat peripheral, long-distance contacts as less certain than close ones.

No bond-type classification. It won't tell you which contacts are hydrogen bonds, salt bridges, or hydrophobic packing — only that atoms are close.

Computed on the asymmetric unit as deposited. If the biologically relevant assembly involves symmetry copies, those contacts won't appear.

An empty result is a real result. No contacts within the cutoff means those two things don't touch at that threshold — that's an answer, not a failure.