This page is dedicated to describe in detail how to use and run mdpocket server. For several reasons, two servers are available to run mdpocket:

• Default server
• Advanced server (Mobyle)

In this page, you will find detailed documentation on how to run mdpocket for both servers (input & output). As the mdpocket methodology is rather complex, we also provide a step by step example of how to analyze a P38 Map kinase sample trajectory.
NOTE : Currently, the webserver is limited to up to 500 snapshots. Gradually during time more and more snapshots will be accepted for submission.

To run mdpocket, you just need a set of PDB snapshots, under the form of a multiple PDB file. Optionally, you can track a specific part (pocket) of the protein by providing a zone of interest to analyse. The following section will describe how to run mdpocket on both server.

Default server

Running mdpocket job

1. Browse (2) your local hard drive to provide a multiple PDB file, or directly paste (1) your PDB file in the text area.
    
   
2. Optionally, enter your email (3) to be notified when your job will be finished.
3. Once your data has been entered properly, you will have to enter the word shown in the captcha image (4). This is to avoid automatic programs to do nasty stuff with our server.
4. You can now run mdpocket by clicking on the Submit button (GO!). Clicking on the clear button will clear all input data you entered

Alternatively, you can test the server, using inhouse demonstration data. To do so, just check the Demo checkbox shown in the picture bellow, and go to step 3 (any input data will be discarded then).

Note that you can give a name to your job using the following text field. The default name contains the service name plus the current date and time.

<Back to top>

Specify a pocket to analyse

MDpocket gives you the opportunity to analyse a specific location in the protein, eg. a pocket of interest. To do so, you need to provide this specific location under the form of a PDB file containing positions of grid points that were selected after a prior mdpocket run without any pocket selection. We recommend the usage of the server to perform this selection :

• Using the server, once you have selected the desired iso-value, you can dowload an archive that contains culstered hot-spot grid points. Based on these files, You can then track the region of your choice by providing the corresponding file to the mdpocket input.
• Using mdpocket output, you can use the pocket grid points file provided by a first mdpocket run and select grid points which will define the zone of interest (using pymol or any other tool).

Once done, you have to provide your data to the server using the following form, which is basically the same as previously described for mendatory mdpocket input:

The output files are described bellow

Intermediate page (job status)

Once you have submitted your data, a new job will be created. You will be redirected to an intermediate page. From there, you just have to wait the end of your job, or bookmark the result page using the adress given in this page. Then, when your job will be finished, you will have a new intermediate page, from where you will have to click on the provided link to obtain the results page described in the output results section.

<Back to top>

Advanced server (Mobyle)

Running a MDpocket job

To run mdpocket using the mobyle server, you need to provide the same input types as described for the default server. The only difference is the interface, and that you have the possibility to retrieve PDBs directly from the Protein Data Bank database.

Here is a short description of the mobyle interface:

1. Provide your multiple PDB information by using the first form labeled Query multiple PDB (PDB), shown in the snapshots bellow. You have 3 possibilities to enter your data, each of these being selected by clicking on the corresponding radio button:
   1. Copy/paste (1): Simply copy and paste your PDB data in the text area.
   2. Database (2): Select your database using the combo box (a), and enter a valid PDB 4 letter code (b).
   3. Local file (b): Select a local file stored on your computer.

1. Optionally, enter the PDB pocket to analyse by using the second form just bellow the first one.
2. You can now run MDpocket by clicking on the Run button (GO!). Clicking on the Reset button will clear all input data you entered
3. The first time you will run mdpocket, you will have to enter a security string. Don't worry, you only have to do this the first time you run mdpocket.

You can also test the server by using inhouse demonstration data. To do so, just check the Demo checkbox shown in the picture bellow, and go to set 3 directly (any input data will be discarded then)

Note that you can also give a name to your job using the following text field.

Specify a pocket to analyse

To specity a specific protein location to analyse as described above, you must fill the following form, which is very similar to that used for mdpocket mendatory input on mobyle:

Intermediate page (job status)

Once you have submitted your data, a new job will be created, and you will be redirected to this new job tab page as shown bellow. From here, you just have to wait for results to appear.

<Back to top>

We use the same output page to display results of mdpocket job for both server. Before describing it, we just describe for both servers, how to reach this page. Don't worry it's really easy.

Getting to the results page

Default server

Just follow the input tutorial, where we describe all intermediates pages from which you will be able to reach the final results page.

Advanced server (Mobyle)

In mobyle, once your job will be finished, you will be redirected to the mobyle results page. This results page is shown in the snapshot on the right. The results page is actually contained in the "Conserved pockets analyse alpha spheres (mdPocket HTML)" labeled scroll pane. To have a more friendly view and reach the output pages described just click on the Full screen view button (1)

You also have the possibility to bookmark the result page using the Bookmark button (2) (enter the bookmark name in the field on left)

Main program output results are provided both in the final results page and in the mobyle interface. This redundancy is comfortable to (i) analyse your results in a specific and integrated web page and (ii) to directly use the mobyle pipelining feature for further analysis.

Results page description: basic mdpocket run (mode 1)

The results can be roughly divided in 3 sections: Output files, Snapshots and Visualisation .

Output files

Several mdpocket output files are provided and described below. You can download all of them.

• MDpocket grid (1): it is the MDpocket output grid that stores density information for each grid point. To learn more about mdpocket methodology and output, go tho the mdpocket method section in the homepage, or read the documentation.
• Pocket grid points (2): This file contains all grid points having 3 or more Voronoi Vertices in the 8A3 volume around the grid point for each snapshot. This output can be used to define a specific zone (a pocket) on which you may want to make further analysis using mdpocket round 2. Additionally, we provide 3 pre-defined PDB files containing grid points up to a given threeshold. Low, medium and high resolution corresponds to low, medium and high isovalues, and thus can be seen as the grid points corresponding to all conserved, and highly conserved cavities.
• Pocket density (3): This file contains the first snapshot, written as PDB file, with the B-factor value matching alpha spheres density nearby this atom. This file allows intuitive, coloured visualisation similar to that of the snapshots described below.

<Back to top>

Snapshots

Two sets of snapshots are provided here: the first set (left picture) represents an alignement of all input structures, and the second set (right picture) represents the first structures surface coloured by alpha spheres density. Here, coulours range from blue (low density = no particular cavity at this place) to red (high density = hot spot = conserved cavity!).

Note that you may obtain such a display by dowloading the output PDB file "Pocket density" described previously, and display it using PyMOL and VMD (color the molecular surface by B-factor). Hopefully, we will allow you to do so directly from here with a new interactive viewer making use of OpenAstex viewer.

<Back to top>

Visualisation

Currently, the visualisation is made using both Jmol and OpenAstex, in which the mdpocket output PDB file is automatically loaded in each viewer (1) (first snapshot).

Using Jmol, you can view the Grid file extracted from mdpocket results. A slider is provided to change isovalue: the highest the isovalue is, the more conserved is the corresponding cavity.
Using OpenAstex, the visualisation is atom-centered. That is, the isovalues have been mapped from the grid to the atoms and transformed to be somehow B-factor-like for coloring purposes. The highest the B-factor is, the more conserved is the cavity associated with atoms.

Both visualisation methods use different metrics. Isovalues will tipically range from 0 to N, with N having no real limitation (depends on the number of snapshots; the slider is limited to 800), while B-factor will range from 0 to 7-8, as it is log-scaled. We are investingating a way to get a common metric, and to merge these visualisation features in a single viewer.

Jmol

Using Jmol, the density grid is loaded along with the protein (1). On the right, you have a set of graphic components to facilitate the viewing. A simple selection box (2) allows you to perform basic changes to the whole system representation (display protein as cartoon, reset view...).

Using the slider on the right, you can change (3) the so called "isovalue". For a given grid point, this isovalue represents the number of alpha spheres seen for all snapshots within a 8A radius. Thus, a high isovalue will display protein cavities which are conserved during the MD, while low isovalues will rather show you every potential transient pockets.

Finally, we give you the opportunity to downlad grid points (PDB format) corresponding to each contigous points that could form a potential pocket(4). To do so, we use an internal clustering procedure from Pymol. You can then use these files as input for the second round of mdpocket (see inputs and ouptuts description), to monitor a specific part of the protein, basically a (potential?) binding site of interest.

OpenAstexViewer

Using OpenAstex, atoms are coloured by B-factor, with the same convention as that used for snapshots, that is, the more an atom is close to red (resp. blue), the more conserved is its corresponding cavity. Only here, the grid isovalue have been mapped on atoms (see accompagning paper for formula) so you can see directely which atom is associated with a conserved zone.

Besides basic visualisation options (2), we give you the possibility to select atoms based on B-factor value (3). By checking the appropriate checkbox, all atoms having B-factor up to the specific value (in the text field) will be selected. If you change this isovalue cutoff, you have to click the update button to update atom selection. 0 is the minimum B-factor value, while the maximum should lie around 7 or 8 (due to log-scaling of isovalues).

Remember that if you know Jmol and OpenAstex, you can access the display popup menu by right clicking on the view.

Results page description: pocket analysis

The pocket analysis will provide 3 output files. Note that there is currently no specific interactive visualisation in the result page, we just provide the result files as is (like in the future mdpocket distribution).

• Pocket vertices (1): This is a pdb file that contains all Voronoi vertices in the selected pocket zone for each snapshot. Each snapshot is handled as separated model (like a NMR structure) and can thus be viewed as MD using PyMOL. Show the surface of the vertices and you can visualize the movement of your pocket. Be careful, VMD does not read this file, as from one snapshot to the other a different number and type of Voronoi vertices can be part of the model.
• Pocket atoms (2): This is a pdb file similar to the previous output, but this time containing all receptor atoms defining the binding pocket.
• Pocket descriptors (3): This file contains all pocket descriptors calculated by MDpocket for each of the input snapshots (1 per line). You can therefore see the evolution of these descriptors, for example the pocket volume... As this output file is a simple text file, it can be easily analyzed using R, gnuplot or other suitable software. An example R output for the pocket volume would be: