Re: [Rdkit-devel] [Rdkit-discuss] RDKit reaction stereochemistry

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

Richard,

I'm moving this from the discuss list to the devel list since it's
getting very implementation specific.

I've got a modified version of your patch applied on a branch
(https://rdkit.svn.sourceforge.net/svnroot/rdkit/branches/ReactionStereochem_23Aug2011)
and it mostly works. The patch was a massive help.

There are still some outstanding things that need to be completed
before calling this "done".
1) You sent a python script with inputs and expected outputs. This
ought to be converted into some unit tests. I'm going to at least get
started on this today.

2)  There's a remaining piece of unsatisfactory behavior that I
haven't been able to come up with a reasonable solution for. It's
demonstrated by the test case test19Issue2050085() in testReaction.cpp
The reaction is one of the RECAP rules:
"[N;!D1;+0](-!@[*:1])-!@[*:2]>>[*][*:1].[*:2][*]"
And the test molecule is: "C1CC1N[C@H](C)F". One of the products of
this reaction is *-[CH](C)F. In a perfect world this would be
*-[C@H](C)F, but I couldn't come up with a way of reliably getting
there. Probably this just needs to be documented.


-greg