[gmx-users] Rigid body optimization

[Matthew Hotchko]

I sent a similar message and no one replied.  I guess that means no one 
knows the answer.  Just in case you missed it, here I go again.  

I am trying to use Gromacs to analyze multiple configurations of a two 
protein complex.  I would like to set the structures with rigid bonds and 
rigid angles, to make both of the proteins rigid bodies.  I tried using 
constraints         =  all-angles
in my md simulation .mdp file, but when I run grompp, I get the following 
error.  
Fatal error: No bond information for bond H1-N or N-H2

What do I need to do in for processing my pdb file with pdb2gmx or what 
else do I need to include in the md.mdp file to get rid of this problem?

Secondly, I would like to do a different MD simulation where only the 
energy of the protein-protein interfaces/interactions is counted.  I have 
two proteins, labled chain E and chain I.  I made an index file for these 
two residues and then tried using the 
energygrp_excl     =  chE chI
energygrps          =  chE  chI
and even adding information for both tau_t and ref_t.  
in my md.mdp, but, I keep getting errors in the process.  What is the 
process from pdb2mgx to mdrun for excluding internal protein energies, so 
I can just look at interface energies.  

Sincerely,
Matthew Hotchko