PELIKAN is a software tool enabling rapid searching of spatial interaction patterns in large collections of protein-ligand complexes. Data from protein-ligand complexes is stored in an SQLite database which can be subject to different search processes. The PELIKAN software package comes with a graphical user interface, providing dialogs to build SQLite databases from any set of protein-ligand complexes and allowing the convenient construction of 3D queries starting from a protein-ligand interface of interest or from scratch. The results of a search are shown in a 3D viewer. The following tasks document different workflows which can be performed with PELIKAN.
The first task shows how to open a database and a filter file in order to perform a quick 3D search with PELIKAN. The database and the filter file used here can be found in the documentation material.
1. Click on File→Open database and select a database file.
Depending on the size of the database, this step can take some seconds.
2. Click on File→Open filter and select a filter file.
If the filter file has been loaded correctly, the details of the filter can be seen in the tabs 'Text/numerical filter' and 'Geometrical filter'.
3. Click on to start the search.
Depending on the database size and the query, this search can take up to minutes. The current step of the search process is always displayed in the upper left corner.
4. Select the 'Results' tab to view the results of the search.
All PDB files detected in this search are shown in the list on the left. By clicking on a PDB code, the list of pockets within this PDB file which fulfill the query and all matches are shown.
5. Click on 'Resort results' to resort the list of results by molecule.
When the results are first presented, they are always sorted by the EC-numer of the protein. Note that for some PDB files, more than one EC-number is annotated. In these cases, the first EC-number is used as category. The result can also be sorted by the unique SMILES of the reference ligand. Here, the category name is the name of one the reference ligands in the respective group.
6. Visualize a match by clicking on a match id.
The complete pocket is displayed. The atoms matching the search points of the query are highlighted. The menu on the lower right corner of the visualization window allows you to show/hide different parts of the pocket.
7. Superimpose different results by selecting another match.
Matches are superimposed using the atoms which match the search points of the query. All displayed matches, their color code and additional information can be found in list on the right side.
8. Click on the icon to export the statistics of the displayed results.
The statistics of your result contain information about all detected PDB files and reference ligands and about all atoms which match the search points of the query, all distance constraints and all angle constraints.
Build a database
This task shows how to build a PELIKAN database out of a collection of protein-ligand complexes.
1. Click on the File→Create new database and enter the name of your new database, its location and the PDB files you want to use.
Protein-ligand complexes can either be added as a complete folder or/and single files can be selected. The calculation of the database usually takes 7s per protein-ligand complex. Afterwards, a triangle descriptor needed for PELIKAN is automatically calculated.
Create a filter with textual and numerical properties
This task shows how to create a query containing textual and numerical filters in PELIKAN.
1. Click on the 'Text/numerical filter' tab.
The list on the left displays all available filter types. In the central view, a list of all textual/numerical filters which are currently part of the query are shown.
2. Choose textual/numerical filters by clicking on the filter name.
In a textual filter, e.g. the filter 'Ligand Name', a string can be entered. This string is used for a case insensitive substring search in the respective entries in the database. Every textual/numerical filter can be including or excluding, as shown in the upper right corner of the filter. The sense of the filter can be changed by clicking on the label.
Create a geometrical filter
This task shows how to create a query containing geometrical filters in PELIKAN. A 3D query can in general be constructed in two ways using PELIKAN. On the one hand, search points, contraints between them and angle constraints can be added from scratch. Moreover, a pocket of interest can be loaded and used to guide the construction of a geoemtrical query.
1. Click on the 'Geometrical filter' tab.
On the left you see three lists. The first list contains all search points which are currently in the 3D query. The second list contains all point-point constraint. The third list contains all angle constrains of the curent query. All details of the constraints can always be adjusted on these lists.
2. Click on the icon to add a search points and define their properties.
Each point filter is expandable by clicking . When a search point is added via the button, all properties of the search point are set to 'Any'. The properties can then be changed using the drop down lists. Every search point has a unique id, shown on the left.
3. Click on the icon to add a point-point constraint. Use the ids of the search points to define for which points this contraints should be checked.
In PELIKAN, there are two different types of point-point constraints: interaction constraints and distance constraints. For the first, 'Use interaction type' has to be selected and an interaction type can be choosen. For the latter, 'Use distance' has to be selected and a distance range can be entered.
4. Click on the icon to add an angle constraint. Use the ids of search points/point-point constraints to define for which elements this contraints should be checked.
Angle constraints can be added between any pair of point-point constraints and directed search points. Search points are directed if they have one of the following interaction types: donor, acceptor, or aromatic. The direction of a point-point constraint is always the direction of their definition (first point id to second point id). The direction of the search points id their interaction direction: h direction (donor), lone pair direction (acceptor), ring normal (aromatic)
5. Click on the icon to clear all lists.
6. Click on 'Load Pocket' to load a pocket for the design of your 3D query.
For the protein file, the file format '.pdb' and '.cif' are supported. For the reference ligand, the file formats '.mol2' and '.sdf' are supported.
7. Click on an atom to add a search point.
If a search point is added by clicking on an atom, all information which can be deduced from the atom is automatically added to the search point, e.g. the element of the amino acid type. Note that interaction types can not be deduced from an atom here.
8. Click on an interaction to add an interaction constraint between two search points.
9. Drack and drop from one search point to another to add a distance constraint between two search points.
The distance ranges which are automatically entered in the distance constraint are the bins which are used by PELIKAN's triangle descriptor. For example, if the measured distance between to atoms is 2.7A, the distance ranges will be 2.5 to 3.5A.
Refine a search
This task shows how to refine a search with PELIKAN. The movie start after a search has been performed.
1. Click on 'Refine results' in the 'Results' tab.
Choose a unique name for your current result set. The current results will be temporarily saved under this name.
2. Change the query and choose the saved results set in the drop down menu. The result can now be refind using all detected result points in the choosen set by clicking on .
3. Alternatively, the search can be refined by using all detected pockets in the choosen set by clicking on .
4. Alternatively, the search can be refined by using all detected pdb files in the choosen set by clicking on .
5. Click on to display the results from the choosen set.