Re: [Cdk-devel] creating SMILES using atom order

[Christoph S. <er...@do...>]

Stefan Kuhn wrote:
> On Friday 22 June 2007 00:43, Rajarshi Guha wrote:
>> Hi, is there a way that the SMILESGenerator can be forced to create a
>> SMILES string using atom ordering rather than generating a canonical
>> SMILES string?
> No. And not just in CDK: If you start with a given atom, you have to continue 
> with the nearest atom and so on. So to have the atoms in a smiles in an 
> arbitrary order is not possible. 

That may not be completely true. You can use "." and bond identifiers
"." (normally used for ring closures) to encode all kind of weird things.

C1CC.CCC1 is n-hexane, as a start.

C1.C12.C23.C34.C45.C56.O6 gives hexanol.

C1.C12.C236.O6.C34.C45.C5 gives hexan-3-ol, to demonstrate a branching.

Try these smiles on http://www.daylight.com/daycgi/depict

Not nice, but it works. Not sure if you can encode a molecule with
arbitrary atom order by this, but it looks like. Haven't thought it through.

Cheers,

Chris

-- 
PD Dr. Christoph Steinbeck
Lecturer in Chemoinformatics
Univ. Tuebingen, WSI-RA, Sand 1, D-72076 Tuebingen, Germany
Phone: (+49/0) 7071-29-78978   Fax: (+49/0) 7071-29-5091

What is man but that lofty spirit - that sense of enterprise.
... Kirk, "I, Mudd," stardate 4513.3..