[gmx-users] Segmentation fault in genion for monolayer-protein system

[Farid Sa'adedin]

Dear Jon

Something is missing from the pdb as your charges are not do not seem 
to add up
correctly.  For GROMACS to work the initial charge should be a whole 
number. Double check the pdb to see if anything is missing.  If not 
just add charges to
your pdb to make it into a whole number.  Then you can use genion to balance
your system.

Farid

Quoting Jon <jon.ellis at utoronto.ca>:

> I am trying to do an energy minimization run of a 148-residue protein 
> attached to a monolayer of 289 short alkyl chains. The protein is 
> attached to the membrane by an amide bond from a lysine to one of the 
> alkys near the centre of the layer (which is ~9.1x9.1 nm^2). I 
> minimized the monolayer in vacuum, then attached the protein by 
> editing the .gro and .top files.
>
> When I run grompp it, it tells me the charge is -2.449600e+01. When I 
> run this in vacuum, it does not minimize properly (I get inf for the 
> forces after only a few steps. I tried adding spc water and ions, but 
> when I run genion, adding 24 NA+, it gives me a segmentation fault 
> immeditely. This reoccurs upon re-grompp-ing, rereunning, and 
> rebooting.
>
> Two questions:
> 1) How many NA+ ions should I add, 24 or 25, for PME to work properly 
> (requires charge =~0)
>
> 2) Has anybody ever seen a segmentation fault in genion, or have any 
> idea what its all about?
>
> Any other suggestions or comments would be welcome.
>
>
> Here's my em.mdp
> ---------------------------------------------
> title		    =  Energy Minimization of protein on surface
> cpp                 =  cpp
> constraints         =  none  ; no bond constraints in EM
> define              =  -DFLEX_SPC
> integrator          =  steep ; steepest descent min to remove bad vdw
> dt                  =  0.002 ; ps
> nsteps              =  50000   ; no. of steps
> nstlist             =  10
> nstenergy           =  2
> ns_type             =  grid
> rlist               =  0.9   ; cut-off distance for list
> coulombtype         =  PME
> rcoulomb            =  0.9   ; cutoff for colomb
> rvdw                =  1.0   ; cutoff for vdw interactions
> fourierspacing      =  0.12
> fourier_nx          =  0
> fourier_ny          =  0
> fourier_nz          =  0
> pme_order           =  7
> ewald_rtol          =  1e-5
> optimize_fft        =  yes
> gen_vel             =  no
> Tcoupl              =  no
> Pcoupl              =  no
> ;	Energy minimizing stuff
> emtol               =  1000
> emstep              =  0.0001
>
> -- 
> ----------
> Jon Ellis
> Institute of Biomaterials and Biomedical Engineering/Dept of Chemistry
> University of Toronto
>
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the www 
> interface or send it to gmx-users-request at gromacs.org.
>

-- 
Farid Sa'adedin
Membrane Biophysics Laboratory
Department of Veterinary Biomedical Sciences
College of Medicine & Veterinary Medicine
Royal (Dick) School of Veterinary Studies
University of Edinburgh
Summerhall
Edinburgh EH9 1QH
SCOTLAND