[Dock-fans] PROBLEMS WITH NCHEMGRID_SA, NCHEMGRID_GB AND ZOU GB/SA SCORING

Scott Brozell sbrozell at rci.rutgers.edu
Tue Feb 28 15:39:01 PST 2017


Hi,

1.
There have been Zou GB/SA Score memory issues over the years.  We have fixed
several including those related to 32 bit vs 64 bit machines.
If you can provide verbose and verbatim details for us to reproduce the
failure then we'll investigate.


2. Do this:
===
cd /opt/UCSF/dock6/install/test/nchemgrid_GB_grid_generation
make check
===
And compare to the comments in install/test/KNOWN_FAILURES, i.e., are these
just insignificant floating point differences ?

scott

On Fri, Feb 24, 2017 at 04:38:23PM +0000, Stellamaris Sotomayor Burneo wrote:
> 
> I am having problems with the Zou GB/SA score. Could you give me a hint about what I can do to solve this problem? 
> 
> I tried the scoring function with two receptors. On my first try the nchemgrid_GB and the nchemgrid_SA were generated without problems. However, by the time I tried to calculate the Zou GB/SA score I obatined this error:
> 
> Initializing Zou GB/SA Score Routines...
> Initializing Grid Score Routines...
>  Reading the energy grid from ../../../grids/grid.nrg
>  Done reading the energy grid.
>  Reading the grid box quantifiers from ../../../grids/grid.bmp
>     size = 2143260
>     spacing = 0.3
>     origin = -26.476,-27.199,32.121
>     span = 126,126,135
>  Done reading the grid box quantifiers.
>  Reading from GB bump grid file nchemgrid_GB/gbsa_zou_grid.bmp
>  Done reading from GB bump grid file.
>  Reading from GB receptor grid file nchemgrid_GB/gbsa_zou_grid.rec
>  Done reading from GB receptor grid file.
>  Reading from SA bump grid file nchemgrid_SA/gbsa_zou_grid.bmp
>  Done reading from SA bump grid file.
>  Reading from SA receptor grid file nchemgrid_SA/gbsa_zou_grid.sas
>  Done reading from SA receptor grid file.
>  Reading from SA receptor info grid file nchemgrid_SA/gbsa_zou_grid.sasmark
>  Done reading from SA receptor info grid file.
> Done Initializing Zou GB/SA Score Routines.
> 
> -----------------------------------
> VERBOSE MOLECULE STATS
> 
>   Number of heavy atoms = 25
>   Number of rotatable bonds = 5
>   Formal Charge = -0.00
>   Molecular Weight = 396.26
>   Heavy Atoms = 25
> 
> GB Calculation
> --------------------
> ligand desolvation:        11.125892           
> receptor desolvation:        31.932617           
> screened electrostatics:    -28.103333          
> total:                14.955177           
> Segmentation fault (core dumped)
> 
> Then I tried with another receptor and I could not get to the step for Zou GB/SA calculation since I had problems to calculate the nchemgrid_SA 
> grids
> 
> I obtained this error:
> 
> Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
> 
> Backtrace for this error:
> #0  0x7F81C39445F7
> #1  0x7F81C3944C3E
> #2  0x7F81C328E2EF
> #3  0x406FB6 in sas_
> #4  0x402268 in MAIN__ at nchemgrid.f:?
> Segmentation fault (core dumped)
> 
> 
> Therefore, I decided to check Dock6.7 with make test and I found problems with nchemgrid_GB and nchemgrid_SA and I found:
> 
> make[4]: Leaving directory '/opt/UCSF/dock6/install/test/gbsa_grid_generation'
> cd ../nchemgrid_GB_grid_generation && make test
> make[4]: Entering directory '/opt/UCSF/dock6/install/test/nchemgrid_GB_grid_generation'
> # Calculate GB grid for Zou GB/SA score
> ../../../bin/nchemgrid_GB > /dev/null
> ../dockdif -w -t 4 OUTCHEM.save OUTCHEM
> diffing OUTCHEM.save with OUTCHEM
> possible FAILURE:  check OUTCHEM.dif
> ==============================================================
> ../dockdif OUTPARM.save OUTPARM
> diffing OUTPARM.save with OUTPARM
> PASSED
> ==============================================================
> make[4]: Leaving directory '/opt/UCSF/dock6/install/test/nchemgrid_GB_grid_generation'
> cd ../nchemgrid_SA_grid_generation && make test
> make[4]: Entering directory '/opt/UCSF/dock6/install/test/nchemgrid_SA_grid_generation'
> # Calculate SA grid for Zou GB/SA score
> ../../../bin/nchemgrid_SA
> ../dockdif -w -t 1 OUTCHEM.save OUTCHEM
> diffing OUTCHEM.save with OUTCHEM
> possible FAILURE:  check OUTCHEM.dif
> 
> Does anybody have a clue about what I can do to solve this problem. 
> 
> I would really appreciate all the help that can be provided!
> 
> I was wondering if this is a Dock bug that has not been solved yet the one that is causing this problems.  


More information about the Dock-fans mailing list