D-EXP002

=Objective= To reproduce the results that were supplied to us by [|Find-A-Drug], and to determine what effect the number of specified centers has in the output numbers and the effect on the computed binding properties performing a 3D site search on THINK. The variables include the number of specified centers and the allowed tolerance. A specified center in THINK is defined by a list of properties including hydrogen bond accepting/donating, acidic/basic properties, aromaticity, and lipophilicity. These center types are outlined in [|Section 7.1 of the THINK Theory Manual 1.25]. According to the Theory Manual, the tolerance is the distance on either side of a center it is allowed to move about, in Angstroms, outlined in [|Section 7.2 of the THINK Theory Manual 1.25]. I'm not 100% sure of this, but if anyone can get a better idea from the documentation, do let us know.

=Procedure=

A flash screencast demonstrating this process will soon be created and uploaded.

This procedure was performed using the [|first 5 molecules] of the 218 molecule library in .sdf format. They were simply copied and pasted into a new Notepad file and saved as this .sdf file. (**NOTE:** .sdf files are a series of .mol files, each separated by $$$$, indicating the end of a section. Make sure that $$$$ is present at the bottom of each molecule, including the very bottom, or you may run into problems. Also, it should not be included at the very top.)

To set the variables
In the File dropdown menu, under Preferences, call forward the Pharmacophores tab. Set the Warp Speed equal to 1, and select the number of centers you want to include in the search (File>Preferences>Pharmacophores, Warp Speed = 1, # Centers).

When performing the 3D search, make sure the Query file is 1p45-q1 by opening it before you bring up this window. Where it says Explore File, click browse and open up the .sdf file in question, the check the box next to Site. Input the tolerance you want to be allowed.

General Procedure
>
 * 1) Open THINK software.
 * 2) Open the [|1p45-q1.pdb] file (File>Open)
 * 1) Set preferences listed above (File>Preferences>Pharmacophores Tab)
 * 2) Open the 3D Search window (Search>3D)
 * 3) Set search options (set Explore File to the .sdf containing the molecules, set query to the protein if it isn't already, set tolerance level, check the box next to Site)
 * 4) Click "Search"
 * 5) It's also helpful if you have the program run in the background, and even hidden. Right click the THINK icon in the system tray and select the Background and/or Hidden options.

Let the search go until it finishes. The search process takes a while (sometimes up to an hour or longer per molecule!), so it's a good idea to run the searches during periods of time you won't be around, or in the background while you're doing other things. This method of docking the DKP's to the protein is able to handle a large number of molecules in .sdf format, allowing a larger number of molecules to tested in a more automated way.

=Results= An Excel document containing a grid of the different settings employed compared to each other, molecule for molecule can be downloaded [|here]. The compared input settings are tolerance, and number of specified centers. The outputs being compared are the G-score, holes, and RMS. The time taken for each run (as listed in the output .sdf file) is also displayed.

=Discussion= The primary numbers we are looking at are the G-scores (the free energy of the enzyme-substrate complex), while the holes and RMS are supplementary for now. A brief explanation, as described by Find-A-Drug through dialogue, is available here.

=Conclusion=

Our results compared to Find-A-Drug's results as a whole contained more holes and a higher free energy value (on average). Some correlations noted were that as the allowed tolerance is increased, the binding free energy and RMS value increases within each group at a set number of centers. The free energy values tended to be closer to Find-A-Drug's initial results with the settings set at 3 or 4 centers, with the tolerance set to 0.5 or 1, or with 2 centers at a tolerance of 0.25. Ways to better this study would be to include the entire library into the searches, and performing them multiple times to check if the data differs in any significant way.

media type="custom" key="7692"