Proteins work because of their three-dimensional shape. Computational models show how a mutation, a drug or a modification changes that shape and, with it, the protein's function. All before taking it to the lab.
Many proteins work like a lock: they have a pocket, the active site, where a specific molecule fits. If the shape of that pocket changes, the key fits worse or no longer fits at all.
Computational structural biology rebuilds that shape in the computer and measures how each piece fits. This answers questions that omics data raise but cannot solve on their own: why a variant disrupts a protein or which compound could act on a target.
Each one answers a different question, and they are combined depending on the project.
The change is introduced into the structure to calculate how it affects stability, shape and contacts with other molecules.
Each molecule is placed in the protein pocket to calculate the position in which it fits and how strongly it binds.
The protein's motion is simulated atom by atom to check whether a binding or a structural change is stable.
These are just a few examples. Almost any question that depends on the shape of a protein, or on how it binds another molecule, can be addressed computationally.
Mutations found in an exome, a panel or a tumor: which contacts are lost and why.
Whether a modification or a mutation closes, opens or distorts the pocket where the protein acts.
How and where a compound binds its target, residue by residue.
Searching approved drugs for those that could bind a new target.
Filtering libraries of millions of compounds down to the few worth testing.
Which components of an extract of known composition could explain its activity.
One compound against proteins of the same family: where it fits and where it does not.
Checking with molecular dynamics that a binding or a structural change holds.
Structure makes the most sense when it starts from an omics result. That is why it is combined with our other services:
The protein in 3D, with the mutation or the compound in place.
Poses, residue-by-residue contacts and comparison between the normal and altered forms.
Which compounds or variants stand out and why.
What the results mean and what is worth validating in the lab.
Computational results do not replace the experiment: they help decide what to test first and with which hypothesis. That saves time, reagents and experiments that were not going to work.
Starting from a protein, a variant or a list of compounds, we assess what can be studied and prepare a proposal.