Overview
RevPocket is Revilico’s pocket search engine. It takes a protein structure as input and identifies all druggable binding sites on that protein — generating a ranked list of pockets with detailed chemistry profiles to guide downstream docking and screening campaigns. RevPocket is typically run after generating or sourcing a protein structure, and before setting up a virtual screening campaign in RevScreen.How RevPocket Works
RevPocket takes a pure protein structure — a PDB file with no ligand or inhibitor present — and runs a cavity detection algorithm to identify all geometrically and chemically favorable binding sites. These pockets are ranked and characterized by their druggability, geometry, and chemical properties. The results help you decide:- Which pocket to target — the active site, an allosteric site, or a cryptic site
- What chemistry to pursue — polarity, hydrophobicity, and surface area drive compound design decisions
- How to set up your docking run — the pocket coordinates feed directly into RevScreen’s docking configuration
Input Requirements
RevPocket requires a PDB file of the apo protein — the protein structure with no inhibitor, cofactor, or ligand bound. This ensures the pocket detection algorithm identifies natural cavities rather than pre-defined binding geometries from a co-crystal. Your protein files are managed in the File Manager (Data Engineering) section of the Revilico OS dashboard.Use the AlphaFold engine to generate a structure if you don’t yet have an experimental PDB file for your target. See the AlphaFold tutorial for the full workflow.
Running a RevPocket Pipeline
Select Your PDB File
In the PDB File Input field, select your apo protein PDB file from the File Manager. Choose the file without any inhibitor present — the pure protein structure.
Review Parameters
All parameters are pre-set to optimized defaults. If you’re an advanced user, click through to the documentation to understand each parameter in detail. For most use cases, the defaults are appropriate.
Reading Your Results
Navigate to the Analysis section and load your completed pipeline to explore the pocket results.Pocket Rankings
RevPocket identifies and ranks up to 10 distinct pockets on your protein. Pockets are displayed as numbered overlays on the 3D structure. Select any individual pocket to isolate it in the 3D viewer and inspect its geometry.Pocket Analytics
Below the 3D viewer, a full analytics table provides chemical characterization for each pocket:| Property | Description |
|---|---|
| Druggability Score | A composite metric estimating how likely this pocket is to bind a small molecule drug |
| Solvent-Accessible Surface Area (SASA) | The surface area exposed to solvent — larger values indicate more open, accessible pockets |
| Polarity | The ratio of polar residues lining the pocket — guides selection of polar vs. hydrophobic compounds |
| Hydrophobicity | The extent of hydrophobic character — high hydrophobicity pockets tend to bind lipophilic fragments well |
Interpreting Results in Biological Context
To decide which pocket to target, cross-reference the RevPocket output with:- Published literature — review papers on known inhibitors or substrates for your target
- Co-crystal structures — if an inhibitor-bound structure exists, identify which RevPocket cavity corresponds to the bound inhibitor pocket
- Substrate biology — understand what molecule naturally engages this protein in its biological pathway, then target the pocket that substrate occupies
Next Steps
With your target pocket defined, you’re ready to move into screening:- AlphaFold — If you still need a protein structure, generate one first
- RevScreen - Static & Flexible Docking — Use your identified pocket coordinates to set up a virtual screening campaign
- RevScreen - Ensemble Docking — Run ensemble docking with MD-sampled conformations for higher-accuracy screening
- Boltz2 Cofolding — Use AI cofolding as an alternative for early-stage pose prediction without pocket pre-definition