Christian,
I believe this was a bug fix where smarts chirality wasn't respected with
dummy atoms. We fixed this during an investigation about r-group
decomposition, basically stereo cores were behaving oddly.
----
Brian Kelley
> On Jun 22, 2016, at 10:30 AM, Kramer, Christian <[email protected]>
> wrote:
>
> Hi Andrew and Greg,
>
> thanks a lot for the quick replies.
>
> I tested Greg's suggested solution, and it works ... but only with the RDKit
> version 2016.03.1. With Version 2015.09.2, I still get the wrong
> stereochemistry after fragmentation (maybe relevant for people working with
> older versions).
>
> Bests,
> Christian
>
>
> Dr. Christian Kramer
> Computer-Aided Drug Design (CADD) Scientist
>
> F. Hoffmann-La Roche Ltd
> Pharma Research and Early Development
> Bldg. 092/2.56
> CH-4070 Basel
>
> Phone +41 61 682 2471
> mailto: [email protected]
>
> Confidentiality Note: This message is intended only for the use of the named
> recipient(s) and may contain confidential and/or proprietary information. If
> you are not the intended recipient, please contact the sender and delete this
> message. Any unauthorized use of the information contained in this message is
> prohibited.
>
>> On Tue, Jun 21, 2016 at 5:54 PM, Andrew Dalke <[email protected]>
>> wrote:
>> On Jun 21, 2016, at 5:26 PM, Greg Landrum wrote:
>> > Because chirality is represented relative to the ordering of the bonds
>> > around an atom, it's pretty difficult to do this if you want to actually
>> > break and add bonds on your own. This would probably be somewhat easier if
>> > there were an RWMol.ReplaceBond() method analogous to the
>> > RWMol.ReplaceAtom() method, but that's not available at the moment.
>> ...
>> > p.s. this all reminds me that there's a long email from Andrew on this
>> > topic that I still haven't worked my way all the way through. <sigh>
>>
>> I pretty much had to give up with working in molecule space and switch to
>> working in SMILES space.
>>
>> That is, I did a SMARTS match or whatever to get the atom to change,
>> backtracked to the original SMILES, which I tokenized to find the
>> corresponding term, then at the token level substituted in the new group.
>>
>> What made it relatively simple was that I wanted to fragment R-groups along
>> non-ring single bonds. In that case, I find the pair of atoms (i, j), along
>> the bond, find the token corresponding to atom j, insert "*.*" before that
>> token, and reparse the modified SMILES.
>>
>> On Jun 21, 2016, at 4:50 PM, Kramer, Christian wrote:
>> > Is there a simple way of preserving chirality during splits on chiral
>> > atoms?
>>
>> To preserve chirality, I had to map from the new molecule space back to the
>> original molecule space, bearing in mind the newly added atom. Then figure
>> out which chiralities were missing in the "*.*"-inserted molecule (since an
>> asymmetric molecule with chirality in the core might have a symmetric core
>> after fragmentation, causing chirality information to disappear), and
>> determine which chirality terms to put back.
>>
>> It's very tricky.
>>
>> I was hoping to present this as the RDKit User's Group Meeting, but I won't
>> be able to make it. :(
>>
>>
>> Andrew
>> [email protected]
>>
>>
>> ------------------------------------------------------------------------------
>> Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San
>> Francisco, CA to explore cutting-edge tech and listen to tech luminaries
>> present their vision of the future. This family event has something for
>> everyone, including kids. Get more information and register today.
>> http://sdm.link/attshape
>> _______________________________________________
>> Rdkit-discuss mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
>
> ------------------------------------------------------------------------------
> Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San
> Francisco, CA to explore cutting-edge tech and listen to tech luminaries
> present their vision of the future. This family event has something for
> everyone, including kids. Get more information and register today.
> http://sdm.link/attshape
> _______________________________________________
> Rdkit-discuss mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
------------------------------------------------------------------------------
Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San
Francisco, CA to explore cutting-edge tech and listen to tech luminaries
present their vision of the future. This family event has something for
everyone, including kids. Get more information and register today.
http://sdm.link/attshape
_______________________________________________
Rdkit-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
