[gmx-users] Possible Bug????

[David van der Spoel]

Matthew Danielson wrote:
> Users,
> I am having a problem when trying to convert a .trr to a .pdb using 
> trjconv.  I think there could there be a bug in the program and the 
> index group is getting confused.  Here is exactly what i am doing:
> Command:
>      trjconv -f eigenvec.trr -s prot_md.gro -o test.pdb
> Output:
> Will write pdb: Protein data bank file
> Select group for output
> Opening library file /usr/local/gromacs/share/gromacs/top/aminoacids.dat
> Group     0 (      System) has 68361 elements
> Group     1 (     Protein) has  7588 elements
> Group     2 (   Protein-H) has  3780 elements
> Group     3 (     C-alpha) has   467 elements
> Group     4 (    Backbone) has  1401 elements
> Group     5 (   MainChain) has  1867 elements
> Group     6 (MainChain+Cb) has  2307 elements
> Group     7 ( MainChain+H) has  2305 elements
> Group     8 (   SideChain) has  5283 elements
> Group     9 ( SideChain-H) has  1913 elements
> Group    10 ( Prot-Masses) has  7588 elements
> Group    11 ( Non-Protein) has 60773 elements
> Group    12 (         FLP) has    30 elements
> Group    13 (         SOL) has 60741 elements
> Group    14 (          Cl) has     2 elements
> Group    15 (       Other) has 60773 elements
> Select a group: 4
> Selected 4: 'Backbone'
> trn version: GMX_trn_file (single precision)
> Reading frame       0 time    0.000
> Back Off! I just backed up test.pdb to ./#test.pdb.1#
> 
> -------------------------------------------------------
> Program trjconv, VERSION 3.3.2
> Source code file: gmx_trjconv.c, line: 994
> 
> Fatal error:
> Index[267] 1403 is larger than the number of atoms in the trajectory 
> file (1401)
> -------------------------------------------------------
> 
> I have been able to get around this problem (really only a work around 
> this bug).  Here is what i have done:
>     editconf -f prot_md.gro -o eigenvec_base.gro -n index.ndx
> output only the Backbone atoms, and then run trjconv with the output 
> from editconf as the template file (-s file).  This then will translate 
> the .trr file to a .pdb.
> 
> Has anyone else seen this, or have any ideas???
> Thanks!
This is the normal behavior. Your eigenvec was made based on backbone 
and therefore has 1401 "coordinates", however in your index file (or 
tpr) there are more atoms, and hence the index for backbone atoms is 
larger than 1401.
Your workaround will create a gro/pdb file, but since the eigenvectors 
are sorted this will not give you normal coordinates. You may want to 
study the options built into the g_anaeig program.




-- 
David van der Spoel, Ph.D., Professor of Biology
Molec. Biophys. group, Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205. Fax: +4618511755.
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se