There are three steps in metaPocket2.0 procedure: calling based methods, meta-pocket site generation and mapping binding residues. The whole working procedure of metaPocket 2.0 is illustrated in Figure 1. In the first step, the given protein structure will be sent to 8 predictors of LIGSITEcs, PASS, Q-SiteFinder, SURFNET, Fpocket, GHECOM, ConCavity and POCASA to identify pocket sites on its surface, all the predictors are called in parallel to save running time. In the second step, the pocket sites identified by these element predictors have different ranking scoring functions, so it is hard to compare and evaluate the predicted pocket sites directly. To make the ranking scores comparable, a z-score is calculated separately for each pocket site in different predictors. Afterwards, only the top three pocket sites in each predictor are taken into further consideration. Therefore, we have a total of 24 pocket sites. Then the pocket sites will be clustered according to their spatial similarity and all the final clusters will be ranked by the total z-score values of them. The final pocket sites are the mass center of the final clusters. The purpose of the third step to identify functional residues around the identified meta-pocket site which could be the potential ligand binding sites on protein surface.
Figure 1. The illustration of the metaPocket 2.0 procedure.
For the detailed description, please refer to our publiction(s):
Bingding Huang (2009), metaPocket: a meta approach to improve protein ligand binding site prediction , Omics, 13(4), 325-330. link,PDF .
Figure 2. The illustration of LIGSITEcs procedure.
Figure 3. The illustration of PASS procedure.
Figure 4. The illustration of SURFNET procedure.
The fpocket core can be resumed by three major steps(Figure 5). During the first step the whole ensemble of alpha spheres is determined from the protein structure. Fpocket returns a pre-filtered collection of spheres. The second step consists in identifying clusters of spheres close together, to identify pockets, and to remove clusters of poor interest. The final step calculates properties from the atoms of the pocket, in order to score each pocket.Reference: Vincent Le Guilloux, Peter Schmidtke and Pierre Tuffery, "Fpocket: An open source platform for ligand pocket detection", BMC Bioinformatics, 2009, 10:168. link.
Figure 5. The illustration of Fpocket procedure.
First project the protein into a 3D grid, the grid width is 0.8 angstrom, then initiate 17 types of different large probes, their radius are 2.0, 2.5, 3.0, 3.5, ..., and 10 angstrom and one small probe S with its radius 1.87 angstrom. Then calculate the multiscale dilation ( ID(X) ), multiscale closing ( or multiscale molecular volume ) ( IC(X) ) and multiscale pocket ( IP(X) ), and the multiscale pocket regions are the binding sites(Figure 6).Reference: Kawabata T. (2010) Detection of multi-scale pockets on protein surfaces using mathematical morphology. Proteins,78, 1195-1121. link.
Figure 6. The illustration of GHECOM procedure.
ConCavity proceeds in three conceptual steps(Figure 7): grid creation, pocket extraction, and residue mapping. First, the structural and evolutionary properties of a given protein are used to create a regular 3D grid surrounding the protein in which the score associated with each grid point represents an estimated likelihood that it overlaps a bound ligand atom (A). Second, groups of contiguous, high-scoring grid points are clustered to extract pockets that adhere to given shape and size constraints (B). Finally, every protein residue is scored with an estimate of how likely it is to bind to a ligand based on its proximity to extracted pockets (C).Reference: Capra JA, Laskowski RA, Thornton JM, Singh M, and Funkhouser TA (2009) Predicting Protein Ligand Binding Sites by Combining Evolutionary Sequence Conservation and 3D Structure. PLoS Comput Biol, 5(12). link.
Figure 5. The illustration of ConCavity procedure.
POCASA uses spheres to identify a layer of probe surface on protein(Figure 8). Those regions between the protein surface and the probe surface are the pocket sites. Besides, by changing the size of probe sphere, pocket sites with different sizes can be detected. POCASA is also a method based on grid since it uses protein atoms to fill the 3D grid system and different radius is used for different atoms. This 3D grid system is divided into a set of slices at the same size and a probe scans from the original point though the slices. Once it encounters a protein grid, it rolls over the protein surface to build the probe surface out of protein surface as mentioned above.Reference: Yu J, Zhou Y, Tanaka I, Yao M (2010) Roll: a new algorithm for the detection of protein pockets and cavities with a rolling probe sphere. Bioinformatics 26: 46-52. link.
Figure 5. The illustration of POCASA procedure.
Zengming Zhang, Yu Li, Biaoyang Lin, Michael Schroeder and Bingding Huang (2011), Identification of cavities on protein surface using multiple computational approaches for drug binding site prediction. Bioinformatics, 27 (15): 2083-2088. link
If you find some bugs of this server, please help us improve metaPocket by reporting bugs to Zengming Zhang, any help from you will be greatly appreciated!
Funding from Klaus Tschira Foundation, MOST China (grant no: 2008DFA11320) and EU 7th Framework Marie Curie Actions IRSES project (grant no: 247097) is kindly acknowledged!