Thanks for the prompt reply Greg; this is what I suspected too!
Regards,
George
On 12 March 2012 17:22, Greg Landrum <[email protected]> wrote:
> Hi George,
>
> On Mon, Mar 12, 2012 at 5:58 PM, George Papadatos <[email protected]>
> wrote:
> >
> > Could anyone please explain this:
> >
> > In [21]: Chem.CanonSmiles('C1=CC=C2C(=C1)NC=S2')
> > Out[21]: 'c1[nH]c2ccccc2s1'
> >
> > In [22]: Chem.MolFromSmiles(Out[21])
> > [16:47:14] Can't kekulize mol
> >
> > In other words, how is it possible that a valid RDKit SMILES output
> fails to
> > be converted to molecule again?
>
> I'm sure the general answer isn't a surprise: it's a bug
>
> It may actually be more than one bug.
>
> The SMILES 'C1=CC=C2C(=C1)NC=S2' probably shouldn't produce a legal
> molecule. It certainly shouldn't produce one with an aromatic ring.
> That's not really a valid/reasonable resonance structure for
> benzothiazole. This would be ok: S1C=NC2=CC=CC=C12 o
>
> The output smiles: 'c1[nH]c2ccccc2s1' is also not a reasonable
> molecule, which the RDKit recognizes when it tries to read it back in.
>
> I'm going to have to think about where the right place to fix this is.
>
> -greg
>
------------------------------------------------------------------------------
Try before you buy = See our experts in action!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-dev2
_______________________________________________
Rdkit-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
