JChemManager Import with Dearomatize standardization rule

Hi,





I tried to import with the latest version of JChemManager import tool a set of aromatized molecules with a custom standardization file so as to get registered non-aromatized molecules.





Although the smiles are indeed in their non-aromatic form, it appears that the sketch of the structures are still aromatized.





is it a normal behavior ?





Thank you





Pierre-Olivier

Hi,





The standardization you provide is used for searching, and not for display.





The structures are always stored in their original format in the cd_structure column, and this is used for display.


That's why you still see them aromatized.





On the other hand:


Dearomatization in this standardizer configuration is a serious mistake.


In fact aromatization should always be present, otherwise the search will not work properly for aromatic query and target structures.


Please read the "Note on aromatic bonds" section of the query guide for details:





http://www.chemaxon.com/jchem/doc/user/Query.html#noteonaromatic





The standardization configuration only effects the standardized form of the structures in the database.


The are stored in the cd_smiles column, and used by the search process - using this column for display or anything else is not recommended.


Of course during the search process the query structures are also standardized with the same configuration that id specified for the database.





For your purposes I suggest running the Standardizer command-line tool first, and import the file after this conversion.


http://www.chemaxon.com/jchem/doc/user/Standardizer.html


(the standardization in the DB can be default or custom, but it should still contain aromatization).








An other option is to standardize the structures right before displaying them (via calling our API). This way each user could have one or more favourite display mode, which does not affect the search.





I hope I could shed some light on things, let us know if you have further questions.





Best regards,





Szilard

Hi Szilard,





OK I understand the standardization affects only the internal representation of the molecule. However could you tell me why the default standardization contains only aromatization and dehydrogenize rules though the "nitro" rule for instance is also mandatory so as to retrieve the same set of molecules when searching as well on [O-][NH+]=O as on O=N=O ? I Know we can add this rule in a specific standardization , but why not include this rule by default if it is mandatory ? What about the other rules ?





Best regards





Pierre-Olivier

Most of the standards might be different company by company, that's why we do not add nitro conversion to the default standardization. For example, the nitro group can be found with the ionic mesomer form as well, but then you must use that format in your queries. Few month ago a professor said during my Standardizer presentation:


"I would have fired you from my exam if you drew nitro in that neutral form."





From a chemist point of view I agreed with him, from the database point of view I still prefer the neutral form. Anyway, you can decide which form will be your standard in your system.

Hi Gyuri,





Thank you for your answer.





Each company may have its specific graphical representations, however the JCHEM internal representation which affects JCHEM search does concern JCHEM only. My question is "do we have to specify" rules or do we have to let JCHEM register molecules with the default standardization rules ? I guess the answer is "we have to" because if we for instance do not specify rules for the "nitro" JCHEM search will fail ? am I right ?





Best Regards





Pierre-Olivier

Yes, you should define your own rules. If you want to find the numerous mesomers, tautomers, salt forms in your database.


Well, you might be right with the mesomers. They should never be handled as different structures. So nitro, azide and other mesomers should be standardized by default. Tautomers and salts can still remain user defined. Hmm...

Ok we agree on the fact that "nitro, azide and other mesomers should be standardized by default".





So why these well known rules are not included in the default JCHEM standardization ? Why let users specify them ?





Best regards





Pierre-Olivier

Most of the standardization actions (especially the transform action) requires Standardizer license. We cannot add these actions to the default standardization configuration of JChem Base.