[gmx-developers] small bug in gmx solvate

[Marc Barbry]

Dear Gromax developers,

First of all, I would like to indicate that I'm a new user of Gromacs, 
therefore some of my comments/questions may be basics. Furthermore, it 
seems that I'm not using the right method to do what is explained below, 
if you could indicate me a better technique to realize this it would 
greatly help me.


Running the gmx solvant program, I detected a bug related to the input. 
When using a basic pdb file for water containing only the coordinates of 
the atoms,

MODEL     1
ATOM      0    O MOL     1       0.000   0.000   0.119  1.00 
0.00           O
ATOM      1    H MOL     1       0.000   0.763  -0.477  1.00 
0.00           H
ATOM      2    H MOL     1       0.000  -0.763  -0.477  1.00 
0.00           H
ENDMDL

Then the program was stuck in a while loop, because the box in 
solvate.cpp was null.


In more details, I'm trying to do the tutorial present in this page
http://www.gromacs.org/@api/deki/files/198/=gmx-tutorial.pdf

You can download the input files with the following link
https://mbarbry.website.fr.to/nextcloud/index.php/s/Wts5ejbyp76ymDp

Thus I have a Lysozyme protein, and I want to add a solvent around it, 
therefore I used the command.

      gmx solvate -cs H2O-nocell.pdb  -cp centered.gro  -p topol.top -o 
solvated.gro

Where H2O-nocell.pdb is the file described above.
In this case the program get stuck in the while loop line ~250 of 
solvant.cpp because box[d][d] = 0.0. Obviously, a check of the box is 
missing here in order to avoid such kind of issue, which are difficult 
to debug.

Second problem, I then added a cell to the H2O.pdb file,

MODEL     1
CRYST1      8.0   10.0   8.0  90.00  90.00  90.00
ORIGX1      1.000000   0.000000  0.000000        0.00000
ORIGX2      0.000000   1.000000  0.000000        0.00000
ORIGX3      0.000000   0.000000  1.000000        0.00000
ATOM      0    O MOL     1       0.000   0.000   0.119  1.00 
0.00           O
ATOM      1    H MOL     1       0.000   0.763  -0.477  1.00 
0.00           H
ATOM      2    H MOL     1       0.000  -0.763  -0.477  1.00 
0.00           H
ENDMDL


The program goes a bit further but then return a memory error

Removed 0 solvent atoms due to solvent-solvent overlap
Removed 120 solvent atoms due to solute-solvent overlap

Program received signal SIGSEGV, Segmentation fault.
0x00007ffff7027dd7 in 
gmx::internal::AnalysisNeighborhoodSearchImpl::~AnalysisNeighborhoodSearchImpl()
     () from /home/marc/.local/gromacs/lib/libgromacs.so.3
(gdb) quit


I guess my input file is not correctly constructed, I'm a newbie with 
pdb files and MD in general. Do you know a good way to generate pdb 
file? At the moment I use Avogadro <http://avogadro.cc/> and ASE 
<https://wiki.fysik.dtu.dk/ase/index.html>, but obviously they were not 
think to be use with Gromacs.

Thanks for your helps,

Best regards,
Marc Barbry
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20180525/58f818f4/attachment.html>