[Dock-fans] Generating New Orientations

[Scott Brozell]

Hi,

On Wed, 1 Apr 2009, Adam Griffith wrote:

> My question is regarding generating ligand orientations with Dock 6.2
> where only bump-grid filtering is used (no energies).
>
> It is my understanding that if I use the same ligand, sphere set, and
> bump grid, and ask for 100 ligand orientations, I will always get the
> same output poses, no matter how many times I run Dock.  Is there any
> way to generate *new* orientations for the same ligand/sphere/bumpgrid
> combination?
>
> For example, I might first ask for 100 orientations, but then decide I
> need to perform additional sampling.  In order to generate 100 *new*
> orientations, I must ask Dock for *200* orientations, the first 100 of
> which will be the same as from the previous run.
>
> Is there any way to avoid redoing the previous work?  Is there a random
> seed that I can set that will give me 100 new orientations each time I
> run with the same settings?  Can I ask Dock for 200 orientations, but
> tell it to skip the first 100 (or however many I asked for in the
> previous run)?

DOCK is deterministic.
The only pseudorandom generator in non-Amber-score DOCK is for the
simplex minimizer.
As far as automated matching, matches are formed best first.
It may be that manual matching may be appropriate 
for your unusual calculation.
http://dock.compbio.ucsf.edu/DOCK_6/dock6_manual.htm#SphereMatching


As far as sampling this post explains the usual DOCK approach:
http://shoichetlab.compbio.ucsf.edu/pipermail/dock-fans/2005-November/000498.html

Scott