[gmx-users] G_SDF if reference atoms are in different residues

[chris.neale at utoronto.ca]

You could use g_spatial instead as that program requires you to do the  
reference calculation yourself with trjconv. I have not tried g_sdf in  
a while, but if you still would rather use that program, then you  
might be able to trjconv your own alignment and then somehow fool  
g_sdf into not doing the alignment, e.g. set all the reference atoms  
as the same atom?

Chris.

-- original message --

Dear Users,

I want to calculate the spatial distribution function of lipid  
headgroups around a protein residue. Because I am using a  
coarse-grained model each 'residue' is a single CG particle. This  
means that when defining the 3 reference groups that define the local  
coordinate axes I have no choice but to choose three groups with  
different residue numbers which I think is causing the program to  
simply not calculate anything and report zeros. G_sdf runs fine when I  
test it on systems in which all the reference groups are in the same  
residue. Is there anyway to post-process my trajectory to fool g_sdf  
into thinking the reference groups all have the same residue number?

This is the message I get at present:
Reading frame       0 time 180000.000
Warning: reference triple (0) will be skipped.
          resnr[1]: 1862, resnr[2]: 1860, resnr[3]: 1861
Last frame       4000 time 240000.000

Back Off! I just backed up refmol.gro to ./#refmol.gro.1#

Wrote reference structure. (3 Atoms)

Number of configuations used for SDF: 0

Mean probability density: 0.000000

Many thanks,
Rob