Structural analysis

Structural Analysis

Sequence to structure to interaction: a staged workflow.

Talk to our computational biology team→All services

A sequence becomes candidate models, each with a confidence value per residue.

A real antibody Fv (PDB 3HFM), coloured by the secondary structure in its coordinates. It shows what a model looks like; nothing was predicted here.

The pipeline

Select a step to see it.

Schematic.

A real antibody Fv (PDB 3HFM), coloured by the secondary structure in its coordinates, to show what a model looks like. It is a deposited structure, not a prediction: no model was run, so no confidence values are shown. Drag to turn.

Drag to turn. Real coordinates, rendered live by IndiskaAI from PDB 1HSG. Any scores shown are illustrative.

Drag to turn. Illustrative simulation, not a result.

Drag to turn. Real coordinates, rendered live by IndiskaAI from PDB 3HFM. Any scores shown are illustrative.

An example summary built from the complex in the previous step (PDB 3HFM). Every figure is computed from its coordinates. It is not a result from a project.

Step 1 of 6

Sequence

The starting input, as supplied.

Step 2 of 6

Structural prediction

Candidate models with per-residue confidence.

Step 3 of 6

Molecular docking

Candidate ways two molecules might associate. A score orders poses; it is not a binding affinity.

Step 4 of 6

Molecular dynamics

Whether the arrangement is maintained in a simulation.

Step 5 of 6

Interaction analysis

Which residues make contact across the interface. On this antibody–lysozyme complex the pairs are walked one at a time, each with its contact type and distance. Proximity in a structure is not proof of an interaction.

Step 6 of 6

Research summary

What the data suggests is worth testing next.

Select another step to see the next.

A staged pipeline, not one technique. Each stage answers something the previous one cannot, and each carries its own assumptions.

Why stages matter

Treating prediction, docking, and simulation as interchangeable is the most common way structural work goes wrong.

From Inputs to Insights

A sequence, or one or more structure files
A partner molecule, if there is one
The question the analysis has to answer

Applications

Drug discoveryBiologicsAntibody engineeringProtein engineeringStructure-based design

Scope & limitations

Prediction produces candidate models; docking and MD ask different questions of them. Neither establishes biological truth on its own; results are computational evidence for experimental follow-up.

Have a structure to understand?

Combine structural modelling, docking and simulation to examine how a molecular system behaves beyond a single static model.

Talk to our computational biology team→Home