[MMTK] PDB producing, amber_atom_charges

[Konrad Hinsen]

> a) I could find in scientific websites files with no default PDB format for
> biomolecules eg
> ALANINE
> -----------
> ATOM      1  N   ALA  0001      -1.053   1.300   0.614
> ATOM      2  CA  ALA  0001      -0.304   0.032   0.746
> ATOM      3  C   ALA  0001       0.770  -0.014  -0.311
...

> Is there a way for example using MMTK.PDB module to create a residue(or a
> peptide for
> more complex case than this) with such format without to change mannually
> H-->1H, C-->CB etc.

At the moment, no. Such a procedure is not so trivial to implement, as it 
would have to match the bonds to the connect information in order to identify 
the atoms.

> b) In the case that I'd like to simulate a more complex organic molecule in
> MMTK or to
> model it in order to use it for a protected group for a peptide.
> I could find coordinates in the similar non-default PDB format:
> NITROANILIDE
> ATOM      1  O           0       2.235   1.226  -0.001  0.00  0.00
> ATOM      2  N           0       1.527  -0.000  -0.000  0.00  0.00
...

> There are 2 subquestions here:
> a) How as in the 1) question to assign partial charges and then fit
> (neutral) them with MMTK
> in order to make with AMBER94 a coarse representation of the organic
> molecule?

In addition to the atom matching problem as before, you need to have Amber 
parameters, in particular partial charges, for your molecule.

> In MMTK database you have only 'water' molecule explicitly defined
> and in groups some simple N,C-terminal org.compounds for proteins but
> is there a more broad option for more complex?

The MMTK database contains everything for which parameters can be obtained 
from the standard Amber parameter files. It is easy to add more molecules 
provided you find the parameters somewhere.

> b)In an earlier email asking about AMBER94 charges you replied that for
> custom cases
> the best is to fit the partial charges to an electrostatic potential
> surface obtained by quantum chemistry.
> Could you be more specific in the quantum utility that does that consistent
> with MMTK-AMBER94? Is there a package, website do you suggest?

There are several packages that can do such calculations (I used Gaussian a 
few years ago). Consistency with Amber 94 requires the use of the right basis 
set, which is explained in the Amber force field publication (which 
unfortunately I don't have at hand now).

> And in more details which is the procedure in MMTK for partial charge
> calculation
> before fitting the charges please?

You use MMTK to construct suitable evaluation points around the molecule, then 
you use a quantum chemistry program to calculate the electrostatic potential 
at these points for the ground state of the Amber-compatible basis set. Then 
you do the fit.
-- 
-------------------------------------------------------------------------------
Konrad Hinsen                            | E-Mail: hinsen at cnrs-orleans.fr
Centre de Biophysique Moleculaire (CNRS) | Tel.: +33-2.38.25.56.24
Rue Charles Sadron                       | Fax:  +33-2.38.63.15.17
45071 Orleans Cedex 2                    | Deutsch/Esperanto/English/
France                                   | Nederlands/Francais
-------------------------------------------------------------------------------