Interface Analysis

Overview

This skill maps protein-protein and protein-ligand interfaces from a coordinate file and quantifies them. It covers contact-residue detection under an explicit distance cutoff, structural epitope and binding-site (ligand-contact) residues, and buried surface area (BSA) as the physical measure of interface size. It also frames the two judgment calls that dominate interface work: which contact definition to use (and always stating the cutoff), and whether an interface observed in the crystal is biological or a packing artifact. The primary tool is Biopython Bio.PDB (NeighborSearch and ShrakeRupley), with freesasa and PDBePISA as complements.

Prerequisites

pip install biopython numpy freesasa

• A structure file (PDB or mmCIF), ideally the biological assembly, not just the asymmetric unit.
• Familiarity with SASA basics (see structural-biology/geometric-analysis).
• PDBePISA (https://www.ebi.ac.uk/pdbe/pisa/) is a web service, no install needed, for the biological-assembly call.

Quick Start

Tell the agent what interface question you have: - "List the residues where chain A contacts chain B in this complex" - "Which residues line the ATP binding site?" - "Map the structural epitope this antibody contacts on the antigen" - "Compute the buried surface area of this dimer interface" - "Is this crystal interface biological or just packing?" - "Find the salt bridges across this interface"

Example Prompts

Contact and interface residues

"Report every residue of chain B within 4.5A of chain A, and tell me which cutoff you used and why." "Give me the contact map between the two chains at an 8A CA-CA definition, and explain what that captures versus a heavy-atom cutoff."

Ligand and epitope mapping

"Identify the binding-site residues within 4A of the bound inhibitor (HETATM LIG)." "This is an antibody-antigen complex; map the structural epitope on the antigen chain and the paratope on the antibody."

Buried surface area

"Compute the buried surface area of the interface between chains A and B and report the per-partner area, keeping the SASA parameters identical across the three calculations." "Compare the interface size of these two dimers computed the same way."

Biological vs crystal interface

"This deposition has two chains in the asymmetric unit; is their interface biological or a crystal-packing contact? Weigh BSA, H-bonds, and what PISA says, and tell me how confident I can be." "Download the biological assembly for this PDB and compute the interface on that, not the deposited asymmetric unit."

What the Agent Will Do

1. Confirm the oligomeric state to analyze, preferring the biological assembly over the asymmetric unit, and download it if needed.
2. Resolve provenance issues first: pick one altloc, decide on waters and crystallization additives, and note missing/disordered residues.
3. Choose and state a contact definition (heavy-atom 4-5A for physical contact, CA-CA 8A for topology) with its rationale.
4. Detect contact residues with NeighborSearch, keeping only cross-partner pairs, and report the interface residue lists per chain.
5. For a ligand or epitope, select the target group (HETATM or partner chain) and collect residues within the cutoff.
6. Quantify the interface as buried surface area, computing complex and isolated-part SASA with identical ShrakeRupley settings.
7. When biological relevance is in question, assemble the probabilistic signals (BSA magnitude, H-bonds/salt bridges, conservation, PISA CSS) and report a hypothesis with a confidence caveat, not a verdict.

Tips

• A contact is defined by an arbitrary cutoff; the residue count changes with it, so always state the cutoff and the atom set (heavy-atom vs CA-CA vs including hydrogens).
• A contact list is not an interface. The physical measure is buried surface area; compute all three SASA terms with identical parameters or the subtraction is meaningless.
• Compute interfaces on the biological assembly. An interface in the asymmetric unit can be pure crystal packing.
• "Biological interface" is a hypothesis. PISA is ~80-90% accurate with known false positives; larger BSA, more H-bonds/salt bridges, complementarity, and conservation raise confidence but do not prove it.
• SASA depends on the 1.4A probe radius and the algorithm; an absolute BSA is only comparable to another BSA computed the same way. Use freesasa for Lee-Richards or full radii control.
• Salt-bridge and H-bond geometric definitions are loose and tool-dependent (distances quoted anywhere from 3.2 to 5.0A); state the exact criteria, and remember angles cannot be checked without modeled hydrogens.
• Double-counting alternate conformations invents impossible contacts; select one altloc before searching.
• Missing loops are disorder, not real chain gaps; reconcile against SEQRES before drawing interface conclusions near a gap.

Related Skills

• geometric-analysis - SASA fundamentals, NeighborSearch, distances that this skill builds on
• structure-io - download the biological assembly before interface analysis
• structure-modification - resolve altlocs and strip waters before contact detection
• structure-navigation - select chains, residues, and HETATM ligands by identity
• structure-validation - check the region is well-fit before trusting an interface
• immunoinformatics/epitope-prediction - structural epitope mapping
• chemoinformatics/virtual-screening - binding-site definition for docking
• alignment/structural-alignment - superpose complexes before comparing interfaces