How we read an interface

The work below is not a customer project. To show how the team’s steps look on real data, we used a publicly available, experimentally solved protein–protein complex.

The question

Through which residues is an interface between two proteins built, and how does that information become a deliverable file?

The input

RCSB PDB 1BRS, biological assembly 1, model 1. Barnase (chain A) and barstar (chain D). The file identity and its SHA-256 are kept in the asset register.

The work

The chains were separated by the label_asym_id field; minimum heavy-atom distances were scanned at a 4.5 Å cutoff. Because the crystal holds three copies of the complex, only biological assembly 1 was used.

The output

43 residue pairs: 19 residues on the barnase side, 16 on the barstar side. The list is produced in the same interface_contacts.csv format we deliver.

Interpretation

The computed interface overlaps the region described for this complex in the literature: residues of the barnase active site (Lys27, Arg59, Arg83, Arg87, His102) appear among the shortest contacts in the list. In other words, the method independently finds a known interface.

What it does not show

This is not a design output: barstar is not a designed candidate but the natural inhibitor of barnase. Nor is the contact list an affinity measurement; it records a geometric fact about the structure. In a design project the same analysis runs on candidate models we generate, and the results are reported as predictions.

Source publicationBuckle AM, Schreiber G, Fersht AR. Biochemistry. 1994;33(30):8878–89.RCSB PDB 1BRS

3D demonstration

The barnase–barstar interface: the workflow shown on real data

This is not a candidate designed by the team. It is a demonstration, on a publicly available and experimentally solved protein–protein complex, of how the workflow steps look.

SourceRCSB PDB 1BRSBuckle AM, Schreiber G, Fersht AR. Biochemistry. 1994;33(30):8878–89.

Ribbon model of the barnase protein — the still image shown before the demonstration starts.

Nothing 3D is downloaded until you press start. The structure file is 390 KB; the viewer library loads separately.

01
Target
What is on screen
A single protein chain: barnase (chain A). The whole surface is visible; no region is marked yet.
In your project
Every project starts by fixing the target structure. Without an experimental structure we prepare a modelled one from the UniProt entry, and the report says that it is a model.
02
Region of interest
What is on screen
19 residues on barnase that come within 4.5 Å of barstar are highlighted. The list was computed from this file, not picked by hand.
In your project
In your project this region is the epitope or hotspot you want engaged. You specify it, or we settle it together through surface analysis.
03
The binding protein
What is on screen
The second chain is added: barstar (chain D), in its own colour. Barstar is not a designed candidate; it is the natural inhibitor of barnase.
In your project
In your project this position is taken by the designed candidate binders. For each candidate you receive a sequence and a predicted complex model, ranked on computational criteria.
04
The interface
What is on screen
Both chains together, with the contacting residues marked on each side. The three closest contacts are drawn as measurement lines with their distances (2.5 / 2.65 / 2.81 Å). No motion here depicts binding taking place; the structure stands as crystallography solved it.
In your project
Interface analysis is the core of the report: contact list, buried surface area, packing quality and a binding energy estimate. These numbers are not affinity measurements; they indicate which candidate is worth testing.
05
Deliverables
What is on screen
The same complex, next to the file actually produced from this demonstration: a 43-row contact list.
In your project
At the end of a project you receive candidate sequences, predicted complex models, a metrics table, the contact list and the method report — including the ranking rationale and the limits.

Limits of this demonstration

This structure is not a design output of the team; it is a publicly available experimental structure.
The demonstration runs no live computation: no docking, design or simulation happens in the browser.
The highlighted region is a geometric contact list, not an indicator of affinity or of a successful design.
The crystal contains three copies of the complex; only biological assembly 1 (chains A and D) is shown here.

A contact was counted where the minimum heavy-atom distance is 4.5 Å or less. Computed on this file, for model 1 and the A–D chain pair.

Terms used on this page

New to the field? Start here. The explanations are deliberately short.

interface: The surface where two proteins touch. Whether a design works shows up largely here.
residue: A single amino acid in the protein chain — one link in the chain.
complex: The structure formed by two or more molecules bound together.
structure: The three-dimensional shape of a protein — a map of where each atom sits.
PDB: The public database of experimentally solved protein structures.
ångström (Å): A length unit at atomic scale — one ten-billionth of a metre; atoms sit a few Å apart.
affinity: How tightly two molecules hold on to each other. It is measured in the lab, not computed.
computational: Work done on a computer. It is not an experiment; it shows which candidate is worth testing.
epitope: The specific patch on the target’s surface where you want the binding to happen.
docking: Computing how a small molecule might sit in a pocket on the target.
UniProt: The public database of protein sequences and their basic annotations.

Describe your target briefly and we will settle the scope together.

The first message never asks for confidential sequences, unpublished structures or sensitive files. Technical detail follows an appropriate confidentiality process.

Let us discuss your project