[gmx-users] possible bug in structure factor calculation in g_rdf?

[Sanku M]

Hi,
   I was using gromacs 4.0.7 to calculate the scattering intensity of polymer solvated in a solvent. But, I found a discrepancy in the normalization of the calculation when considering only a subset
  of  the entire system ( e.g . only solute from the entire solution) :
This is what I found:

 I  first tried to calculate the scattering intensity of  *only* the polymer ( i.e the solute) using trajectory file of the entire solution and .tpr file of the entire solution using an index file as shown below

g_rdf_4mpi -s topol -f traj -noxvgr -n -sq sq_polymer_only  << EOF
 6
 EOF
( Here index-group 6 is the entire polymer).

But, then I  generated  a .tpr file *only* for the solute ( i.e no solvent) and used trjconv to generate the .xtc file *only* for the solute 

trjconv_4mpi -s topol.tpr -f traj.xtc -n index.ndx -o traj_polymer_only.xtc


Now, If I try to calculate the scattering intensity, this time using newly generated .xtc file *only* containing the solute and the new .tpr file *only* containing the solute as follows:
g_rdf_4mpi -s topol_polymer_only.tpr -f traj_polymer_only.xtc -noxvgr  -sq sq_polymer_only1 


I get much higher scattering intensity compared to the earlier case ( where  trajectory of the entire solution was used and a index file was provided to consider the solute only for calculation purpose), however the peak position is same for both cases. I think, possibly, there might be a bug which does the normalization for the entire solution in stead of the subset of the system if asked to. 
Any help in clarifying it will be helpful.


Sanku
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110812/0c886772/attachment.html>