[gmx-users] index files for trjconv -sub

Fabian Paul Fabian.PAul at mpikg.mpg.de
Mon May 14 10:59:02 CEST 2012

Dear Gromacs users,

I'm new to molecular dynamics and I face some problems with the analysis of my
I intend to use the trjconv utility to extract frames from my trajectories. Each
trajectory can contain several frames that belong to different clusters. To do
this I create an .ndx file with the following structure
[ output_file_name_1 ]
index1 index2 ... indexN
[ output_file_name_2  ]
index1 index2 ... indexM
Then I pass this to trjconv in the following fashion:

trjconv -s my_topology.gro -f my_trajectory.xtc -sub my_index_file.ndx

I expect trjconv to split the trajectory into several xtcs, where the section
names in the index file are taken as the file names for the output. And each
newly created xtc will only contain these frames of the original trajectory that
where given as indices in the respective section of the .ndx file.
So is my understanding of the trjconv functionality correct so far???
It tured out that this works most of the time, but fails sometimes. Sometimes
the last frame that was specified in the indices (indexN or indexM) is not
written. This also goes wrong (sometimes) when there is just a single index in
a section. 

What puzzeles me even more is that I get the error:
| Double entries in block structure. Item 3 is in blocks 0 and 0.
| Cannot make unambiguous inverse block.
for an innocuous looking index file with the content:
[ test ]
2 6 4 0

Maybe there is something wrong with the 0. Doesn't the frame count start from 0?
Ah, and my gromacs version is 4.5.5.
Fabian Paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20120514/9d483c1b/attachment.html>

More information about the gromacs.org_gmx-users mailing list