1. Target definition
All tools in the ProteinsPlus server work on a structure of a
protein or a nucleic acid. In the first step, you need to define this structure either
by selecting a structure form the Protein Data Bank (PDB) via its PDB code (e.g. 1a99)
or upload any other structure in PDB format. Additional ligand molecules can be provided
in SDF format.
Click on image to jump to that page.
2. Tool selection
Afterwards you can select the tool you are interested in.
You can choose GeoMine either by selecting it from the drop-down menu or by following
the link under the respective tool description.
General mouse/keyboard controls for the 3D viewer
include the rotation/translation/zooming of the scene, inspection of atom/bond infos and
execution of structural measurements.
Click on image to jump to that page.
3. Set Structure visualization for 3D query selection
GeoMine performs textual, numerical and 3D searches
in protein-ligand pockets of the entire PDB dataset. Possible use cases are the
search for ligands for a given pocket or off-target proteins for a given ligand.
Prepare the visualization in the 3D viewer to facilitate the 3D query selection:
E.g. select and focus a precalculated pocket entry from the "Pockets" tab,
set all "Representation Options" below the 3D viewer
to "Off" in order to visualize only the pocket, use "Toggle Viewer Size"
(try fullscreen) and set "Background" to "White".
Besides the 3D query, a GeoMine search might also contain a PDB subselection,
which restricts the search to a user-defined set of PDB codes, and
filters for textual/numerical properties of any component of the
protein-ligand complex, e.g. the enzyme class or the hydrophobicity of a pocket.
All search types can be used independently or in combination.
Press "Search" in any of the query tabs to start the GeoMine search.
Click on image to jump to that page.
4.1 3D query selection
The 3D viewer allows the construction of queries from scratch and based on
the 3D representation of an existing protein-ligand complex or
its pockets. You can select the 3D query in any
structure, the most convenient way is to select it
in a precalculated pocket from the "Pockets" tab.
The precalculated pockets are processed by Protoss, DoGSiteScorer and GeoMine during calculation.
Empty pockets with a volume less than 100 cubic angstrom are excluded and
their number is limited to twice the number of protein chains.
If DogSite could not calculate a pocket for a ligand, the radius pocket (6.5 angstrom)
of the ligand is added.
The pockets contain water molecules, ions, ligands, other small molecules and interactions of
the following types:
H-Bond (blue), Ionic (magenta), Metal (yellow),
Pi-Cation (green), Pi-Pi (cyan).
Furthermore, selectable GeoMine features (binding site surface atoms,
aromatic ring centers and normals, end/mid alpha carbon atoms and directions of
helices and strands)
are visualized.
The pocket can be switched on and off guiding the 3D query design.
The 3D query is visualized in the 3D viewer and
in tables for further modification of chemical and geometrical properties.
Main properties like the element name of a point are automatically set.
The automatic loading of all other, more discriminative properties
can be activated with the checkbox "Point selection with all properties".
For a valid 3D query, all points must be part of at least one point to point constraint.
Click on image to jump to that page.
4.2 3D query selection
Choose a selection mode in order to select a 3D query in the 3D viewer.
Point: Define points by clicking on (surface) atoms and point features
(aromatic ring center, secondary structure mid/end alpha carbon atoms).
Surface atoms are represented by non transparent big spheres.
Undefined points can be added clicking on an unoccupied position in the 3D viewer.
These points can subsequently be translated in the 3D viewer
to the desired position, if the point is not connected by point-point constraints.
In order to specify an undefined point as aromatic ring center, its "Element"
must be set to "Any" and its "Type" to "Aromatic".
The selected points can be further specified by different attributes.
Distance/Interaction: Point-point constraints (distances and interactions)
can be defined by clicking on a pair of points.
They can also be added directly by clicking on atoms, aromatic ring centers and
secondary structure points. The corresponding points are automatically added.
The interaction type (e.g. Ionic) and point types (e.g. Cation/Anion) will be automatically set,
if an interaction is available between the points. The default tolerance value is 0.5 A
and preset for each added distance range.
Distance length and tolerance can be modified by the user.
Angle: Angle constraints can be defined between connected point-point
constraints clicking on these two. Also ring normals of aromatic ring centers
and directions of secondary structure features
can be used for angle definition. The preset tolerance is 15 degree.
Angle size and tolerance can be modified by the user.
Hide the precalculated pocket in order to facilitate
the selection of point-point constraints and angles between already set points.
For achieving a short search time, the following rules are useful:
(1) Set as many properties of 3D query as possible, e.g. define a protein point
(backbone or sidechain);
(2) Choose rarely occurring attributes for the 3D query, e.g. prefer the selection of N
instead of C backbone atoms;
(3) Add as many textual/numerical properties as possbile;
(4) Add as many distance and interaction constraints as possible;
(5) Choose small distance tolerances in close proximities.
Click on image to jump to that page.
5. Textual / numerical property filter selection
The numerical and textual properties can be added to a table. The filters
can be negated and a PDB must fulfill all defined filters to be a match. For the
textual filters, a list of properties can be defined. Here a matching PDB must fulfill
only one of these properties. You can inspect a filter description of the filter that
is currently selected in the dropdown by hovering over the info icon next to it.
Click on image to jump to that page.
6. PDB codes selection
Put PDB codes (case insensitive) into the respective text box to limit
the search on a PDB subselection. Please separate the PDB codes by newlines.
If the text box is empty, the complete database will be searched. Using the result
refinement feature, all resulting PDB codes that matched the search criteria are
loaded into this text box for further refinement of the search and results.
Click on image to jump to that page.
6. XML query
The currently selected query can be converted
to XML which can then be used for the REST API.
Click on image to jump to that page.
7. Result
An search process of successive query modifications and
result analyses can be performed using the refinement and history functionality.
The results can be refined continuing the search in all found PDB codes
and with the current 3D query as a new starting point.
The results history allows browsing through prior and successive
search results obtained by refinement.
GeoMine produces a list of all matching PDBs and respective pocket matches.
The first 150 results can be displayed as superimpositions onto the 3D query in the viewer
for visual analysis with different representation options.
E.g. set "Ligand Representation" to "None"
and the "Protein Representation" to "Pocket" in order to compare pockets visually.
The table can be searched, e.g. with respect to the PDB title, and its entries sorted, e.g.
by their enzyme classes.
Besides the visual analysis of hits, the pockets of the first 150 results
can be downloaded as .pdb files, the table content as .json and .csv files and common structural
aspects of the complete results list can be inspected by means of a statistics report.
Click on image to jump to that page. If you want to access
your results later, you can copy a link to the results.
