[Dock-fans] Re: FW: Serious Issues!!! (John Irwin)

Demetri Moustakas moustakas at gmail.com
Fri Apr 29 12:22:51 PDT 2005

Dear Shuang,

This is not a memory leak, but rather reflects the random number generator 
used by DOCK during energy minimization.  If you re-order the molecules in 
the input file, the value of the random number sequences will be different, 
resulting in different minimization values.  This is equivilant to changing 
the random number seed.

However, this is an indication that you are not sufficiently sampling in 
your system.  If you request more orientations, you will find the same 
minimum energy position regardless of the random number seed.  I recommend 
you test this by changing the random seed, and trying different numbers of 
orientations.  You should look for the orientational sampling level that 
gives similar results regardless of the random seed.


>Message: 1
>Date: Thu, 28 Apr 2005 22:50:11 -0700
>From: "John Irwin" <jji at cgl.ucsf.edu>
>Subject: [Dock-fans] FW: Serious Issues!!!
>To: <dock-fans at docking.org>
>Cc: 'shuang luo' <shuang260 at yahoo.com>
>Message-ID: <200504290550.j3T5oJlp2767758 at uracil.cgl.ucsf.edu>
>Content-Type: text/plain; charset="us-ascii"
>Hi - Here is another mail from Shuang Luo suggesting some kind of memory
>leak (?) that results in different results depending on the order of the
>ligands in the input to DOCK5.2. Anyone?
>   _____
>From: shuang luo [mailto:shuang260 at yahoo.com]
>Sent: Thursday, April 28, 2005 1:19 PM
>To: John Irwin
>Subject: Serious Issues!!!
>Dear Dr.Irwin:
>Here is one "serious issue" that need your help.
>Is there any particular parameter to set when docking a library? I mean
>besides those like, multiple_ligands etc.
>I found that dock5.2 show different results to ligand regarding to where it
>is in the concatenated file. Let me explain in this way, I have 3 ligands,
>ligA, ligB and ligC, then concated into one big
>mol2 file with the order of ligA-ligB-libC. The docking pose given by
>dock5.2 of ligA is the same when dock ligA individually. But the docking
>pose is different for ligB and ligC compared to those docking individually.
>In other words, if I concatenated the mol2 file in the order of
>ligB-ligA-ligC, the pose of ligA is different from the pose when docking
>ligA individually. In some case the difference is quite big.
>You can try that for dihydrofolate reductase, 1DF7.pdb, 1DG5.pdb and
>1DG7.pdb. This also happen to all the ricin complex I docked.
>So obviously, there is some setting that should be reseted after each round
>of docking
>when docking a library of ligands. I wonder what are they? or maybe it is a
>Really looking forward to hearing from you.
>With best regards

More information about the Dock-fans mailing list