The Molecular Graphics Laboratory Forum

AutoDock, AutoLigand, MGLTools, Vina, PyRx and more.
It is currently Sun Apr 22, 2018 9:57 pm

All times are UTC




Post new topic Reply to topic  [ 5 posts ] 
Author Message
 Post subject: Variable Results
PostPosted: Wed Jul 08, 2009 12:41 am 
Offline
Millimolar User
Millimolar User
User avatar

Joined: Tue Jul 07, 2009 11:15 pm
Posts: 3
Hello,

I'm using Vina to dock some molecules against yeast HSP82. Before docking the unknowns, I want to make sure I can replicate a set a known co-crystals.

In a nutshell, I'm getting values sometimes very close to the co-crystal ligand (RMSD ~ 0.5A), but the result seems to be highly dependent on the random seed. I've tried setting exhaustiveness up to 200, but if the random seed is a certain value, the increased exhaustiveness does not seem to help.

A random seed that generates a poor result is -696703708 (-6.8 kcal/mol , RMS = 7.866)
A random seed that generates a good result is 105310060 (-9.5 kcal/mol , RMS = 0.517)

In general, is it best to run a docking a number of times with a smaller exhaustiveness value to find a good result?

Thanks,
Jeremiah

Code:
receptor = 1A4H_apo.pdbqt
ligand  = 1A4H_lig.pdbqt
size_x = 57
size_y = 49
size_z = 51

seed = -696703708

center_x = 22.3235
center_y = -6.994
center_z = -2.76
exhaustiveness = 25
out = 1A4H_apo_1A4H_lig.pdbqt
log = 1A4H_apo_1A4H_lig_vina.log


Attachments:
1A4H_lig.pdbqt [4.1 KiB]
Downloaded 247 times
1A4H_apo.pdbqt [162.14 KiB]
Downloaded 281 times
Top
 Profile  
 
 Post subject: Re: Variable Results
PostPosted: Wed Jul 08, 2009 5:56 pm 
Offline
MGL member
MGL member
User avatar

Joined: Mon Oct 20, 2008 9:39 pm
Posts: 202
Jeremiah,

It doesn't make sense to try to "choose" the seed, but it does make sense to make the search space smaller, when possible.

Your search space is large enough to include all of the receptor and a lot of the surrounding space. This reduces the probability of finding the global minimum at the default exhaustiveness (but can be overcome by setting the exhaustiveness higher, much higher in your case).

I chose a different search space with size 20x20x20, centered at (13, -3, -3) (see attached picture), and re-ran your docking at the default exhaustiveness (8). This gave me the correct conformation 9 times out of 10.

Good luck,
Oleg


Attachments:
1A4H_box.png
1A4H_box.png [ 190.44 KiB | Viewed 6221 times ]

_________________
If you don't get your question answered here, consider posting it on the AutoDock mailing list instead of this forum. Please do not email or PM me with requests for help.
Top
 Profile  
 
 Post subject: Re: Variable Results
PostPosted: Wed Jul 08, 2009 6:13 pm 
Offline
Millimolar User
Millimolar User
User avatar

Joined: Tue Jul 07, 2009 11:15 pm
Posts: 3
Hi Oleg,

Thank you for the insight.

Some of the ligands I'll be docking will not necessarily bind to the same pocket, which is why I included the entire protein in the grid.

For the blind docking scenario, should 5-10 dockings be performed and the lowest energy binders considered the best result?

Thanks,
Jeremiah


Top
 Profile  
 
 Post subject: Re: Variable Results
PostPosted: Wed Jul 08, 2009 11:38 pm 
Offline
MGL member
MGL member
User avatar

Joined: Mon Oct 20, 2008 9:39 pm
Posts: 202
Quote:
should 5-10 dockings be performed and the lowest energy binders considered the best result?

This wouldn't be very different from increasing the exhaustiveness 5-10-fold (there are some subtle differences that you can figure out by reading the paper). For your ligand-receptor complex and search space, it does not look like exhaustiveness=80 would be sufficient though.

Quote:
Some of the ligands I'll be docking will not necessarily bind to the same pocket

One normally targets a particular binding site in virtual screening. While some ligands can bind outside of the intended binding site, do you care enough about them to justify a very significant trade-off in speed? This should probably guide your choices.

_________________
If you don't get your question answered here, consider posting it on the AutoDock mailing list instead of this forum. Please do not email or PM me with requests for help.


Top
 Profile  
 
 Post subject: Re: Variable Results
PostPosted: Fri Jul 10, 2009 5:26 pm 
Offline
Millimolar User
Millimolar User
User avatar

Joined: Tue Jul 07, 2009 11:15 pm
Posts: 3
oleg_trott wrote:
Quote:
should 5-10 dockings be performed and the lowest energy binders considered the best result?

This wouldn't be very different from increasing the exhaustiveness 5-10-fold (there are some subtle differences that you can figure out by reading the paper). For your ligand-receptor complex and search space, it does not look like exhaustiveness=80 would be sufficient though.



You are, of course, correct. The random seed that did not find the binding orientation at exhaustiveness 25 or 200, did find it at 2000.

Thanks for your help.
Jeremiah


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 5 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
POWERED_BY
Translated by MaĆ«l Soucaze © 2009 phpBB.fr