[OTDev] Are there some sample dataset services available ?

Jörg Kurt Wegner joerg.wegner at web.de
Tue Feb 16 10:18:00 CET 2010


Nina,

 

>> Finally, still, in theory, mapping hashed InChIKeys for "identical"
structures is possible, 
>> whatever identical means. 

>Interesting, do you have a reference?



Not really, the “colored graph canonicalization” for molecular compounds is
quite controversial due to a lack of open standards.

Anyway, two references and more reading could be [ig94] and
<http://www.opensmiles.org/> http://www.opensmiles.org/, but as said many
times by OpenEye, in practice, consistency might be better than accuracy (to
an undefined goal).

InChIKeys are the best I have seen so far, if all identifiers of all
possible tautomeric and protomeric forms are mapped together in some
relationship.

 
>>If not, multiple input structures, aka "identical" InChIKeys, should get
used. 

>I am afraid we don't have a clear decision currently how to proceed in this
case, but fortunately, the project is still running :)  

>The dataset service supports multiple structures per compound, we would
need to properly flag structures and 

>agree how these are processed by calculation services.



I am a paranoid industry person, so I need “no tox flags” or the “minimum
percentage of tox flags” for all potential protomeric/tautomeric input
structures, since for me all is about risk minimization.

I would always challenge a “single” input structure, if it was not reported
that this is the “correct form”, pH, and tissue we will observe TOX in. How
do I know? 

 

Cheers, Joerg

  

 

@ARTICLE{ig94,

  author = {W. D. Ihlenfeldt and J. Gasteiger},

  title = {{H}ash {C}odes for the {I}dentification and {C}lassification of
{M}olecular

                {S}tructure {E}lements},

  journal = {J. Comp. Chem.},

  year = {1994},

  volume = {15},

  pages = {793--813},

  contents = {hash code, hierarchical algorithm, Augmented Connectivity
Molecular

                Formula (ACMF), Cahn-Ingold-Prelog (CIP), stereochemistry,
resonance

                structure},

  doi = {10.1002/jcc.540150802},

  groupsearch = {0},

  topics = {hash code, hierarchical algorithm, Augmented Connectivity
Molecular

                Formula (ACMF), Cahn-Ingold-Prelog (CIP), stereochemistry,
resonance

                structure}

}




More information about the Development mailing list