From: Ashar Malik (
Date: Tue Dec 09 2014 - 16:50:08 CST

In your MD - you were already using SHAKE for your hydrogen atoms. So the
current representation of your system has emerged from that. I guess the
question you should be asking is that in the MD ( which is basically MM)
the area of your interest that you now want to resolve, i.e. where your
ligand is interacting how many hydrogens are at the interface, if there
aren't many, you could delete hydrogens, and add them all over again using
psfgen effectively creating a new PSF of your system. However, what would
be better is that you write a simple code that iterates over your PDB file
and changes atom types.

e.g. look here:

or there is an another program that lists table for equivalance here:

a little before the half way point on the above webpage.

This would help you keep your atom's positions.

Also - there should be programs, i think, which already achieve this.

If you are interetsed in deleting and guessing positions again using
psfgen, that should be "OK" but you might want to wait till you hear back
from an expert in the list ---

Good luck,


On Wed, Dec 10, 2014 at 4:39 AM, Kshatresh Dutta Dubey <>

> Dear Users,
> I have amber generated trajectories and I am willing to take some
> snapshots for QM/MM calculation using chemshell that runs fine with charmm
> psf and pdb file. But the different nomenclature of hydrogen for charmm and
> amber is creating the problem from direct use of amber generated pdbs. Is
> it correct if I save the coordinates without H and then re-adding them in
> psfgen and going further for QM/MM calculations? Please suggest me about
> the appropriateness of this method.
> Regards
> Kshatresh