¶ Introduction
The Pocket Browser allows the user to explore all pockets and ligands of a given protein structure. Both ligand-binding sites and ligand-agnostic pockets are identified and characterized by fpocket (related paper) when a structure is registered in 3decision. Each pocket is characterized based on the properties such as druggability, polarity, volume, hydrophobicity, and charge.
There are two types of pockets:
- explicit pockets, when the pocket is defined by the presence of a ligand (represented by squares in the legacy version)
- implicit pockets are pockets detected by fpocket (represented by circles in the legacy version)
The Pocket Browser table is interactive with the 3D structural information. Click on the pockets to navigate in the structure through the different pockets.
¶ Access
You can access it from the Workspace, clicking on the navigation bar on the top: it is the third in the drop-down menu.
¶ Column types
¶ Pocket Number
That is the identifier of the pocket. It is an internal value, defined by 3decision.
¶ Druggability
The druggability score is a numerical value between 0 and 1 associated to each pocket using a logistic function. This scores intends to assess the likeliness of the pocket to bind a small drug-like molecule. A low score indicates that drug like molecules are likely to not bind to this pocket. A druggability score at 0.5 (the threshold) indicates that binding of prodrugs or drug-like molecules can be possible. 1 indicates that binding of drug-like molecules is very likely.
In the paper Understanding and Predicting Druggability. A High-Throughput Method for Detection of Drug Binding Sites, an alternative scoring function is introduced in fpocket allowing you to assess the likelihood of a binding site to bind small druglike molecules. Since the publication, the score has been retrained and performance improved. The theoretical basis of the score is currently in the lengthy process of scientific publication.
Roughly, if you get a druggability score of 0 or close to 0 it is predicted not-druggable with drug-like molecules. If the score is above 0.5 there might be a chance to find drug-like molecules. The druggability assessment is done using some of the pocket descriptors extracted by fpocket. A low druggability score DOES BY NO MEANS indicate that no molecule binds to a pocket. I.E. a peptide binding site will bind peptides, but peptides won't necessarily be considered as drug-like molecules.
A druglike ligand is defined as a compound with Molecular Weight between 150 and 2000 Dalton, at least one ring, and a single bond fraction < 0.95 (single bond fraction = # single bonds / total # of bonds in the molecule).
¶ Ligands
This column contains the three-letter codes of ligands (from RCSB) that are in the corresponding pocket.
¶ Pockets Residues
This column contains the list of residues that are lining each pocket. The naming convention is the following: (Chain name(residue one-letter code:residue number))
¶ Ligand Delimited Pocket
This column contains two different values:
- 1, if it is an explicit pocket, meaning that the pocket is defined by a ligand
- 0, if it is an implicit pocket, detected by fpocket
¶ Volume in A3
As indicated by the name, this descriptor tries to evaluate the volume in Angström cubed of a binding pocket, using a Monte Carlo algorithm. It calculates the full volume occupied by all alpha spheres in a given pocket.
¶ Surface Area in A2
That is the surface area of the pocket in Angström squared.
¶ Polar Surface Area in A2
That is the polar surface area of the pocket in Angström squared.
This descriptor provides an estimation of the polar surface area of the pocket based on information about the receptor atoms. The method used to calculate the area only provides an approximation, but should be good enough to get some rather relevant estimates.
¶ Apolar Surface Area in A2
That is the apolar surface area of the pocket in Angström square.
See polar surface area in the previous point, only for apolar atoms.
¶ Flexibility
The flexibility score is a numerical value between 0 and 1 associated to each pocket.
Please handle this score with a lot of care for native crystal structures. This score is based on the mean B-factor of all atoms of the binding pocket (atoms that are contacted by at least one alpha sphere). As the B-factor does not necessarily reflect flexibility in crystal structures, this score is somehow abusive. However, one could imagine performing molecular dynamics or other in order to determine relative flexibility of atoms and store this information in the B-factor column of the PDB file format. This descriptor is normalized with other pockets of the same protein.
¶ Hydrophobicity
This descriptor is based on a residue based hydrophobicity scale published by Monera & al. in the Journal of Protein Science 1, 319329 (1995). For all residues implicated in the binding site, the mean hydrophobicity score is calculated and is used as a descriptor for the whole pocket. Each residue is evaluated only once.
¶ Charge
This descriptor is the total charge of residues lining the pocket. Individual charge by residue can be retrieved in the following table.
| Charge | Value | Residues |
|---|---|---|
| Neutral | 0 | A C F G I L M N P Q S T V W Y |
| Charge + | +1 | H K R |
| Charge - | -1 | D E |
¶ Bulkiness
This descriptor tends to measure the density and “buriedness” of a pocket. It is nothing else than the mean value of all alpha sphere pair to pair distances in the binding pocket. Thus, a small value indicates a rather big compactness of the binding pocket and thus a rather buried pocket. Larger values give an indication of more extended and exposed cavities.
¶ Polarity
On contrary to hydrophobicity, this descriptor tries to measure the hydrophilicity character of a binding pocket. To each residue of the binding pocket, a polarity score is assigned. The final polarity score is the mean of all polarity scores of all residues in the binding pocket. This is extremely approximative, so should not be overestimated. Each residue is evaluated only once.
¶ Reference
Tables for properties calculation
Fpocket paper
Fpocket github
Druggability paper