gmx trjconv functionality should check for mismatch between the tpr and the trajectory
this happens in both 5.1.4 and 2016.3
system1 and system2 have different number of atoms and different molecules.
But gmx trjconv doesn't check if the trajectory that is being read matches with the tpr file that is given in the -s option.
the program continues to write a wrong trajectory output that matches the tpr file.
for instance say i had a trajectory of 200 benzene and i supplied a tpr of 300 waters, trjconv still writes an output.
I checked few other commands like rms, rmsf, they do issue an error when the analysis starts that there is a topology mismatch.
Not sure if the feature is meant to be implemented this way, but there can be a check in place whether the topology in the tpr matches with the trajectory in all cases.
#1 Updated by Berk Hess over 2 years ago
- Status changed from New to Rejected
trjconv can not check for this. We like to allow e.g. reading a trajecory of a protein only for a system of protein+water. Since most trajectory formats do not have any atom or molecule information, we can not perform any checks (except for the that #atoms in the trajectory should be <= than that of the topology, which we do not seem to check).
#3 Updated by Mark Abraham over 2 years ago
- Tracker changed from Bug to Task
- Subject changed from gmx trjconv doesn't report an error if there is a mismatch between the tpr and the trajectory to gmx trjconv functionality should check for mismatch between the tpr and the trajectory
- Status changed from Rejected to Accepted
There's definitely too many things being done in trjconv, and some of them are probably amenable to such sanity checks. But such checks make sense as part of the "analysis" framework (and perhaps exist there already). I'll convert this to a future task, and link it to the other "trjconv sucks" issues.