Re: [Rdkit-discuss] UFF/MMFF atom types

[Greg L. <gre...@gm...>]

Hi Michal,


On Mon, Nov 4, 2013 at 11:45 AM, Michal Krompiec
<mic...@gm...>wrote:

> Hello,
> Is Se defined in UFF and/or MMFF94? Apparently, molecules with
> selenophene moieties don't optimize in RDKit, and a warning appears in
> the log: UFFTYPER: Unrecognized atom type: Se2+2
>

Right. UFF Parameters are there for sp3 Se ("Se3+2"), but not for the sp2
version.

There are no MMFF94 parameters for Se.

Is it possible to define/modify the force field by hand? (for example,
> use the parametrs of S for Se)
>

If you are working from C++, you can provide UFF parameters when you build
the force field, but it's not currently possible to do so from Python. It's
probably possible to add an option to the python UFF code to allow you to
provide a new atom type (or to over-ride parameters for an existing atom
type); I'd have to look into that.
In the meantime, the quickest thing you could do would be to modify
$RDBASE/Code/ForceField/UFF/Params.cpp to add the atom type you want and
then to rebuild the RDKit.

I guess that adding new atom types to MMFF94S is considerably more complex.
Here one could imagine providing an interface to explicitly set the type of
an atom to another existing atom type. Paolo would have to let us know how
much work that is.

-greg