[Dock-fans] Dock6.2 installation - possible failures judgment

[Scott Brozell]

Hi,

On Wed, 26 Mar 2008, Ben Keshet wrote:

> 
> Thanks for looking at my test output.  The first file is indeed 
> 'secondary2_conformers.mol2.dif' as you corrected.  Where the 
> differences are significant, as in 
> 'primary_secondary_score/secondary7.dockout.dif, you wrote that you 
> think that these difference are likely due to the grids being generated 
> on different machines.  Do you think that this is only a "scaling" 
> difference? That is, do you think that overall - docking positions and 
> ranks, would generally be qualitatively similar across different machines? 

Yes.

> I visually inspected 'secondary7_secondary_conformers.mol2' and the 
> original 'secondary7_secondary_conformers.mol2' as well as the 
> 'secondary7_secondary_scored.mol2' files (using Chimera) and they look 
> different, although some conformers look alike (not the best one). 
> 
> On a related note - the same possible failures occurred when I installed 
> dock6.1 a while ago.  I ran part of the test_suite available online 
> (first 23 structures), which is validated for dock5.4, with dock6.1.  
> The rigid scores were off by ~5 and the flexible by ~2.  The rmsd of the 
> scored ligands from the 'ligand.mol2' structure in each set of 
> receptor-ligand are 2-3 angstrom with only one docked ligand being 
> completely off. 
> 
> Taken together, it looks to me like despite these failure my machine 
> docked the test_suite successfully for the most part.  If you have any 
> other measures I could use to validate the installation and performance 
> of dock on my machine I'd be happy to use them. 

Not really; I notice this
Parallel Processing Test Set
http://dock.compbio.ucsf.edu/Test_Sets/index.htm

Scott

> Thanks,
> Ben
> 
> Scott Brozell wrote:
> > Hi,
> >
> > On Mon, 17 Mar 2008, Ben Keshet wrote:
> >
> >   
> >> I installed Dock6.2 and have several 'possible failures' after running 
> >> 'make test'.  I inspected the output of 'make check' and most failures 
> >> were insignificant differences in tails of floating point.  However, 
> >> differences in the scores in secondary_conformers.dif and 
> >> secondary_scores.dif(2,4, and 7) seem more significant.  secondary7 
> >> (primary + secondary with minimization) has huge differences in the grid 
> >> and contact scores.
> >> Some of the .dif files are pasted below. I am running dock on a PC with 
> >> cygwin.  I would greatly appreciate if anyone experienced users could 
> >> help me in judging the severity of these failures or suggest how to fix 
> >> them.
> >>
> >> ./primary_secondary_score/secondary2_secondary_conformers.mol2.dif
...