Thanks Geoff,
I took some time to really examine the SMARTS language today and it's not
so bad - I panicked when I saw things like aliphatic/aromatic.
Instead of writing a parser, I am going to try and use a bunch of if/else
clauses. Would it be logically sound to proceed from bottom of the list and
work my way up? That way I take care of the most specific cases first, and
then proceed to general cases. I am assuming here that more specific cases
farther down the list (e.g. atom [#6X2](#*)-*#*) take precedence over
[#6X2] and such even though an atom may fit both categories.
One question remains then - what is a good way to determine whether an atom
is aromatic?
Thanks,
- Eric
On Thu, Dec 22, 2011 at 12:27 PM, Geoff Hutchison
<ge...@geoffhutchison.net>wrote:
> You don't have to use SMARTS -- you could program the atom types with a
> set of if/else clauses. Essentially all of MMFF94 atom typing is done this
> way in Open Babel. It's probably more tedious, though.
>
> > Sorry, but my question remains as to how can element is assigned its
> atom type based solely on bond connectivities.
>
>
> Molecular mechanics implementations like GAFF define a set of atom types
> determined by humans. So if you have the valence bond structure of a
> molecule (i.e., atomic numbers, bonded neighbors, and bond types = double,
> triple, etc.) you should have everything needed for GAFF.
>
> The only "other" things I see in the list of GAFF atom types are ring
> membership and aromaticity. You can, most definitely, assign these based on
> bond connectivities.
>
> So I guess my question is -- what more do you think one would need beyond
> atomic numbers and bond connectivities? In the organic chemist's view of
> the world, that's essentially everything.
>
> Thanks,
> -Geoff
------------------------------------------------------------------------------
Write once. Port to many.
Get the SDK and tools to simplify cross-platform app development. Create
new or port existing apps to sell to consumers worldwide. Explore the
Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join
http://p.sf.net/sfu/intel-appdev
_______________________________________________
OpenBabel-discuss mailing list
OpenBabel-discuss@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openbabel-discuss
