Hi, I believe I've found a bug in the new code that deals with reactions that have chirality specified for untagged product atoms. Consider the following:
> rxn = AllChem.ReactionFromSmarts("[C:1].[C:2]>>C1C[C@@H](C[C:1])CC[C@
@H](C[C:2])1")
> m1 = Chem.MolFromSmiles('FC')
> m2 = Chem.MolFromSmiles('BrC')
> p = rxn.RunReactants((m1,m2))[0][0]
> Chem.SanitizeMol(p)
rdkit.Chem.rdmolops.SanitizeFlags.SANITIZE_NONE
> Chem.MolToSmiles(p, isomericSmiles=True)
'FCC[C@@H]1CC[C@@H](CCBr)CC1'
The output looks about right, the [C@@H]s are both still [C@@H] but whereas
before they were both being approached from around the ring now the
canonicalisation has us approaching one from outside the ring. Both
extensions from the ring should be towards and if I convert the product
part of the above reaction to a png I get:
[image: Inline images 1]
but in the output one is towards and the other is away:
[image: Inline images 2]
Note that I can work around this, if I specify my reaction as
[C:1].[C:2]>>[C:1]C[C@H]1CC[C@@H](C[C:2])CC1 thus apeing the atom ordering
of the product RDKit will give me I get the chirality I want, at least in
my test cases so far.
Yours,
Toby Wright
--
InhibOx Ltd
<<inline: ReactionProduct.png>>
<<inline: MadeProduct.png>>
------------------------------------------------------------------------------
_______________________________________________ Rdkit-discuss mailing list Rdkit-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
