idTarget

Documentation

Ligand:
The idTarget web server requires an input ligand file and a set of receptor structure files for the so-called inverse screening. The user can upload a ligand file with the popular molecular file formats (cif, pdb, mol2, or pdbqt). Internally, idTarget uses the “ pdbqt” file format that contains the information of atomic charges and atom types. For a detail description of the pdbqt file format, please refer to this page. The user can select the Gasteiger, AM1-BCC or RESP charge model for the uploaded ligand. Otherwise, the default atomic charges will be calculated by the AM1-BCC charge model.

Protein sets:
idTarget employs a contraction-and-expansion strategy to perform target screening efficiently. A screening job will start with a contracted list of representative pockets for the protein-ligand complexes in Protein Data Bank. Next, the targets with lower docking energy are kept and idTarget will expansively screen the homology of those. In addition to the contraction-and-expansion strategy, idTarget provides two kinds of modes for searching binding poses. The scanning mode represents that the normal docking jobs are performed for each protein structure. In the Fast mode, the ligand was mapped to the binding site and locally minimized instead of performing docking when a homology has been screened.

2. Docking and Scoring

MEDock ¹ is used to perform docking against the entire PDB or a user-edited set of PDB entries. No matter which docking program will be used, the docking poses will be locally minimized and rescored by the robust AutoDock4 scoring function. Finally, the binding free energy calculated with the robust AutoDock4 scoring functionin ² will be reported.

To analyze the results of target screening, the affinity profile of each binding site is constructed. The affinity profile of a given pocket is modeled by a Gaussian function, whose width is determined by the range of the predicted binding affinities of the complexes of different ligands in the same binding pocket of the same protein. A Z-score of ligand j to a given protein pocket i is calculated based on above profile using the following equation:

where E_ij is the dock energy of ligand j docking to the protein pocket i. E^c represents the energy directly evaluated with the crystal pose. E_i and sd_i are the center and the width of the affinity profile of protein pocket i, respectively. A large negative Z value signifies an important target of the query compound.

3. The divide-and-conquer docking approach

Since the binding site is usually unknown for new targets, we adopt a divide-and-conquer docking approach for efficient blind-docking to search the entire receptor surface instead of only a part of surface or just the active site. This allows the search for allosteric binding sites. The divide-and-conquer docking approach is a simple idea which takes the advantage of the parallel computing facilities and reduces the searching time by limiting the size of grids. Therefore, the divide-and-conquer docking approach not only increases the efficiency but also the probability of finding the global minimum than the traditional docking approach. The following figures llustrate our divide-and-conquer docking approach:

Firstly, a box with the buffer length of10Å to each boundary of the receptor was drawn (white). This big box was subsequently divided into smaller boxes (cyan; the cyan dots are grid box centers) where the size box was dynamically determined according to the size of query ligand. The suitably overlapped boxes (red and green dots are the grid box centers of overlapped boxes) will be taken into count. If a grid box is far away the receptor (no atom is within 1.42*length of the grid box to the center), it will be eliminated to reduce the computational cost. Finally, the boxes shown above are retained.

4. Job Submission

Before submitting an idTarget job, you could upload the ligand file in most of the popular structural file formats (pdb, mol2, etc.) or the pdbqt file with partial charges assigned. If you have a ligand with determined protonated state, please check "Yes, no further adding polar hydrogens is needed". You can simply draw a ligand molecule with the Marvin Sketch applet embedded in idTarget web server.

As default setting, the larger scale screening is performed with idTarget dataset, or you can also query any protein with known structure by editing a PDB ID list.

Finally, you may choose the docking methods of idTarget. If your e-mail address is provided, the result will be e-mailed to you as soon as the calculation is finished. In the period of calculation, you may see some part of finished jobs immediately. The time to finish a idTarget task depends on the number of docking targets and the server loading. It usually takes about one day to finish a idTarget task when the idTarget dataset is used.

5. Results

You may check the server loading in the queue page. You may also check the progress and view current result of a submitted job by entering the Task ID. The estimated binding free energy will be sorted right after the docking of each target protein finished. Users can inspect the binding poses and manipulate the resulting 3D structures interactively via the OpenAstexViewer applet embedded in the result page.

By clicking the button, the user can show or hide the detailed information about this idTarget task. If no email provided upon job submission, the user have the second chance to provide email by clicking the button. If for any reason the user wants to cancel the idTarget task, the user can terminal this idTarget task by clicking the button. The two functions, and , are only available when the idTarget task is running or in queue.

While inspecting the binding poses and the resulting 3D structures interactively via the OpenAstexViewer applet in the result page, the user can change the protein reprentation simply by checking the box. By default, the protein will be shown in cartoon representation while the queried ligand will be represented in sticks and transparent molecular surface in magenta as well. The user can check the box to show the original complexed ligand in the PDB crystal structure for quick comparison. The original ligand will be shown in yellow sticks and yellow molecular surface if it is available in the PDB entry.

Please refer to our example page for illustration of results in deatil.

1. Chang, D., et al. (2005) MEDock：a web server for efficient prediction of ligand binding sites based on a novel optimization algorithm Nucl. Acids Res., 33, W233-W238.
2. J.C. Wang, J.H. Lin, C.M. Chen, A.L. Perryman, A.J. Olson (2011) Robust scoring functions for protein-ligand interactions with quantum chemical charge models, J. Chem. Inf. Model. 51, 2528-2537.