Irrelevant trans not ignored?

Hi guys,





We noticed recently that the following two structures are considered different by JChem Base:





OC(=O)\C=C(\C)C





and





OC(=O)C=C(C)C





but when the structures are rendered they are correctly portrayed in the same way. Just to be perfectly clear: the explicit trans in the first structure is superfluous because there are two methyls on one side of the double bond. What am I missing?





Thanks for your help,





Moses





Note: I did check this with the latest JChem 5.2

Hi Moses,





I recommend to use the global stereo model.





See: http://www.chemaxon.com/jchem/doc/user/query_stereochemistry.html#stereomodels








It should work with the comprehensive model as well (which we plan to make the default), but for some reason it does not. We will check it.











Best regards,








Szabolcs

Hi,

The default stereo model is currently the local stereo model. As you can read in the query guide local stereo model doesn't check the symmetric nature of the query or target. Hence your stereo example won't match the non-stereo one.

The reason behind this behaviour is, that local stereo model is suited for substructure search, where a symmetric query can match to an asymmetric target. In this case a symmetric stereo query requires a stereo target for the matching e.g.

your query: C\C(C)=C\C(O)=O

target: CC\C(C)=C/C(O)=O

The same target without the stereo information isn't found.

 

In comprehensive model your stereo specific and non-stereo structures are matching, because it doesn't require stereo target if the target is symmetrical.

Bye,

Robert

Hello,

In comprehensive model the stereo specific structure matches the non-specific for full and substructure searchtypes, but didn't match it in duplicate search. We corrected this and the fix will be released with version 5.2.1 within the next two weeks.

Please note that currently the default model is the local stereo model, so you have to change it to global (or to comprehensive after 5.2.1) in order to match such structures.

Bye,

Robert

 

Thanks, this is helpful. How do you set the stereo model for inserting structures with the UpdateHandler?

Hello Krishna,

Unfortunately you cannot set the search parameters with updateHandler.

Are you interrested in such a feature? We are considering the implementation of this.

Until that you can make a work-around:

Enable duplicate entries to be imported. (UpdateHandler.setDuplicateFiltering(boolean))

Open the structures one-by one, perform a search, and modify the search option by setting the stereo model.

e.g.

JChemSearch jcs = new ....;

initialize it

jcs.getSearchOptions().setStereoModel(SearchConstants.STEREO_MODEL_COMPREHENSIVE);

 

Bye,

Robert

 

Please see more about duplicate searching using the Java API in the developers guide:

http://www.chemaxon.com/jchem/doc/guide/search/index.html#duplicate

 

Or if you are using the cartridge, you should use jc_compare:

http://www.chemaxon.com/jchem/doc/guide/cartridge/cartapi.html#jc_compare

 

Best regards,

Szabolcs

rwagner wrote:

Hello Krishna, Unfortunately you cannot set the search parameters with updateHandler. Are you interrested in such a feature? We are considering the implementation of this.

Thanks rwagner & Szablocs. Yes, we would definitely use this feature. The duplicate filtering option of the update handler seems useless if an extra search must be performed anyway prior to inserting a new structure.

Best,

Krishna

 

Hello,

We decided to implement it. Most probably this function will be realized as a new table option, where the user can set the desired stereo-model.

The next major release is due about october-november.

Bye,

Robert

 

Hi Robert,

Great news, thanks,


Moses

Hello,

 

We decided to implement the ignorance of symmetric strereo centers in case of duplicate search in a different way.

General table option is not suitable, because the most users might wish to use local or comprehensive  stereo model for substructure searches but global stereo model for duplicate searches. A table option would mean that for the given table only that stereo model is applied per default regardless of search type.

So we decided to change the default stereo model in the following way:

If the stereo model is defined explicitly, then this will be used,

Otherwise the default settings apply:

for duplicate/full search and non-query tables the global stereo model is applied. (this applies for inserting)

for duplicate/full search and query tables the local stereo model is applied.

for other search types  - comprehensive model is used.

 

These changes might solve your problem with DB inserting, because for duplicate searching the global stereo model is used, which takes into account the symmetric nature of the stereo center.

However for substructure search it will still be possible to search with a symmetric query for asymmetric, stereo specific structures because the comprehensive model will be used.

 

This modification can be introduced in the next major version (minor versions are suited for bug fixes.)

 

Will these modifications fulfill your requirements?

Bye,

Robert

 

Hi Robert,

Yes, that will work perfectly. Good point about the table option/substructure searching.

Moses

Hi,

Since JChem 5.3 the default stereo model is the comprehensive model (for some query and markush tables the local model):

https://www.chemaxon.com/jchem/doc/user/query_stereochemistry.html#stereomodels

So the symmetric targets are matched by the stereo queries.

Regards,

Robert