What does PDBSiteScan do?

How does PDBSiteScan work?

·        PDBSiteScan provides automated search of three-dimensional (3D) protein fragments similar in structure to known functional sites. A collection of known sites we designated as PDBSite was set up by automated processing of the PDB database, using the data on site localization in the SITE field. Additionally, protein-protein interaction sites were generated by analysis of contact residues in heterocomplexes. We accepted a residue as contact, if it had at least three atoms whose distance from any atoms of the partner chain was smaller than 5 Å.

·        The algorithm developed is based on exhaustion of all the possible combinations of protein positions to be compared with the site. PDBSiteScan accomplishes the following steps. At the first step, the amino acids of a protein part and the site are compared. If they are identical, their 3D structures are compared at the second step. If the maximum distance mismatch (MDM) of superimposed structures is smaller that the user-specified value, the fragment examined is added to the list of the results. Our approach considers only atoms N, Ca , C that define the spatial orientation of residues.

Using the PDBSiteScan Interface

Input for PDBSiteScan

Step 1. Pick a structure to scan

·        Input a filename of the tertiary structure of a protein under study into the text window. The tertiary structure should be in the PDB format. You can upload it by clicking the "Browse" button.

·        Input ID chain of the protein to be analyzed.

Step2. Specify the threshold value of the MDM

Decision making, whether or not to include the canning results for the potential functional sites, structurally similar to the real, in the results, depends on the MDM threshold value. The requirements for the structural similarity of the potential and real sites become less stringent with decreasing threshold MDM values. At high values, potential sites very different from the real are not discarded by the program, thereby producing overprediction of functional sites; at low values, the potential sites even structurally similar to the real, can be discarded thereby producing underprediction of functional sites. The optimum MDM threshold values vary in the narrow 1.0 – 2.5 A range.

Step 3. Choose a type of site to scan

PDBSiteScan searches active sites, binding sites and posttranslational modification sites in 3D protein structure. To search sites of a particular type, click checkbox next to the site name. When a site group is marked with a checkbox “All in a group”, all the sites in the group become predictable by PDBSiteScan.

Step 4. Submit query

Click "Scan" button at the bottom of the page.

Step 5. View results

·        To obtain a structural alignment, choose the site of interest from the list of the identified sites by placing a checkbox next to the site name. Then, click “Download structure alignment as PDB file”.

·        To view the PDBSite database entry, corresponding to the identified site, click ID of the given site.

PDBSiteScan Output

·        The output contains information on the superimposition for each site–protein fragment pair. It also includes a unique PDBSite identifier of the site; the protein PDB ID from which the site was extracted; site description; the MDM and Root Mean Square Deviation (RMSD) values for the identified protein fragment; structural alignment data. Each result is linked to the complete information on the site in PDBSite.

·        PDBSiteScan provides structure site–protein alignment as a PDB file. This allows visualization of the structure alignment by the popular software, such as Chime and RasMol.


Let us consider the application of PDBSiteScan, using a recognition of a catalytic site in the 1ELV protein (the HYDROLASE family) as an example.

This figure demonstrates the input data for PDBSiteScan.

This figure demonstrates the result of PDBSiteScan operation.

Comments and questions are welcome to Vladimir Ivanisenko