[gmx-developers] request for trjconv -pbc pull_pbcatom

[chris.neale at utoronto.ca]

Hello,

When applying the pull code and pulling something toward the center of  
mass of an aggregate of highly dynamic molecules (e.g. detergent  
micelle), the pull_pbcatom0 selection can be quite important, and yet  
it is not necessarily obvious if the box is sufficiently large, even  
for a smart selection of pull_pbcatom0. One could obviously just make  
the box larger, but it would be nice to avoid that.

As an alternative, it would be nice to have an option to trjconv that  
functions exactly as the new pull code and outputs the periodic image  
of each atom that is closest to the selected pull_pbcatom. One could  
then view a long-time trajectory and empirically determine if the box  
is big enough for the study to be undertaken, or, more likely, run  
g_dist over all pairs and automate anomaly detection.

I suspect that g_principal could be utilized here to determine cases  
in which the box is much too small, but I can't think of a tool that  
will detect small and infrequent problems.

It is also unclear if the px or pf .xvg files from the pull code could  
be used to detect such problems, but I think that these would not  
catch the cases in which only a few atoms were put in the "wrong"  
image for COM determination.

Thanks,
Chris.