[MMTK] Reading in of AMBER related files

Konrad Hinsen hinsen at cnrs-orleans.fr
Wed Apr 4 07:42:36 UTC 2007

On 03.04.2007, at 19:52, Mark Williamson wrote:

> Can MMTK parse in a AMBER parm topology file
> (http://amber.scripps.edu/formats.html) and then say evaluate the  
> energy
> of a structure in a corresponding AMBER coordinate file?

MMTK has a parser for Amber parameter files, which is in fact used  
every time an instance of Amber99ForceField is created. But there is  
no parser for Amber topology files.

> It seems the AMBER topology is inferred from the PDB file upon  
> adding of
> the Object to the universe.
> Where do the electrostatic charges come from?

The topology is taken from MMTK's chemical database, which consists  
of a bunch of files in the directory MMTK/Database. There is one file  
for each atom type (the chemical elements plus a few "united atom"  
types), one for each group (sidechains, peptide group, etc.), one for  
each non-chain molecule (e.g. water), and there can be database files  
for proteins and nucleic acids as well, though they are more usually  
built from a PDB file. The database files were initially constructed  
by conversion from the Amber 94 topology files, but lots of  
information was added later by hand. They contain among others Amber  
atom type and charge information.

There is an alternative way to create groups and molecules from  
Python code, by using a MoleculeFactory. However, this approach is  
not used (currently at least) for reading in PDB files.

> How does MMTK deal with non-standard resides?

For a non-standard residue you have to write a new database file. It  
would normally be located in your personal database (~/.mmtk/ 
Database), but the use of databases can be customiz:ed through an  
environment variable, in particular to allow sharing of databases in  
a research group or via the Web.

> Also, in a given universe, is it possible to list every dihedral term
> (with their parameters) within the force field that the universe is
> using? I suspect due to the OO nature of MMTK this would be quite easy
> to do.

Yes. It happens that I added an example script for that feature just  


Konrad Hinsen
Centre de Biophysique Moléculaire, CNRS Orléans
Synchrotron Soleil - Division Expériences
Saint Aubin - BP 48
91192 Gif sur Yvette Cedex, France
Tel. +33-1 69 35 97 15
E-Mail: hinsen at cnrs-orleans.fr

More information about the mmtk mailing list