pdb(topology/trajectory)reader should not be based on libgmx

[GoogleCodeExporter]

pbd files are handy to use as topology base, which happens most likely without 
gromacs.

Original issue reported on code.google.com by christop...@gmail.com on 6 Sep 2013 at 9:01

[GoogleCodeExporter]

Original comment by christop...@gmail.com on 6 Sep 2013 at 9:22

[GoogleCodeExporter]

I think this is one of the first things to do on the way to NAMD support.

Original comment by christop...@gmail.com on 4 Dec 2013 at 2:26

• Changed title: pdb(topology/trajectory)reader should not be based on libgmx

[GoogleCodeExporter]

With the upcoming gromacs 5.0 release this get even more urgent.

Sikandar can you have a look at this? The  xyzreader and the pdbwriter should 
be a good starting point.

Original comment by jungh...@votca.org on 12 Dec 2013 at 7:10

[GoogleCodeExporter]

Sure, Christoph, I will start looking into it.

Original comment by symasha...@gmail.com on 12 Dec 2013 at 7:14

[GoogleCodeExporter]

The (stupid) pdb,xyz, and gro parsers already exist in votca-ctp. If you want 
to make readers out of it i can send you the code.

Original comment by andrienko@googlemail.com on 12 Dec 2013 at 9:05

[GoogleCodeExporter]

This seems ctp_map uses CSG::TrajectoryWriter, CSG::TrajectoryReader and 
CSG::TopologyReader.

Original comment by christop...@gmail.com on 12 Dec 2013 at 9:27

• Changed state: Accepted

[GoogleCodeExporter]

yes, ctp_map relies on CSG reader and writer, but many objects in the libctp 
have their own readers/writers, mostly for test purposes. this is a 50 lines 
code if you don't want to have it 100% user-safe. There was no sense to change 
ctp_map (we considered this) since the Topology reader is essential there - 
rewriting it would be a pain.

Original comment by andrienko@googlemail.com on 12 Dec 2013 at 9:47

[GoogleCodeExporter]

PDB2Map::readPDB from src/libctp/tools/pdb2map.h looks good. I copied some 
parts to a new pdbwriter. Right now this code will only be compile with 
-DWITH_GMX=OFF and it fail from multi-frame pdb files. 

Original comment by christop...@gmail.com on 12 Dec 2013 at 11:54

[GoogleCodeExporter]

I fixed reading of multi-frame pdb files. When reading a pdb trajectory written 
by csg_map, I realized that our pdb writer is actually not 100% correct, it 
writes the charges in the wrong column, so I didn't implement reading charges 
in the reader either.

Denis, do you make use of the charges in csg_map?

Original comment by christop...@gmail.com on 16 Dec 2013 at 4:37

[GoogleCodeExporter]

Nope, we read charges from mps files (distributed multipoles) - force-field
charges make no sense for evaluating electrostatic energies only. No need
to have them in the pdb.

Original comment by andrienko@googlemail.com on 16 Dec 2013 at 5:49

[GoogleCodeExporter]

Great, in csg charges aren't use either and there is a warning to make the user 
aware.

Sikandar/Victor, please review my code and if you feel it is good enough, 
please close this issue. 

Original comment by christop...@gmail.com on 16 Dec 2013 at 5:58