User 81a38f9467
20-06-2014 17:06:46
Duplicate searches should find mesomeric duplicates., since such duplicates are merely different representations of a single structure.
I am using JChem Cartridge 6.2.1 and the searches
select * from compound where jc_compare(compound.structure, '[Br-].CN1C=C[N+]2=C1CSC2','t:d') = 1;
and
select * from compound where jc_compare(compound.structure, '[Br-].C[N+]1=C2CSCN2C=C1','t:d') = 1;
should both return the structures 1 & 2 in the example set (attached). However, each of these searches only finds one of the two compounds.
Please fix.
P.S. I posted a related bug here: http://www.chemaxon.com/forum/viewtopic.php?p=57036#57036
ChemAxon d4fff15f08
23-06-2014 13:57:49
Hi Tim,
This a problem indeed. Do you have any kind of standardization on your table? The Mesomerize action (http://www.chemaxon.com/jchem/doc/user/standardizer_actions.html#mesomerize ) is designed to do the standardization for the mesomers so you'll be able to find them during duplicate searches (in this case both 1&2). Unfortunately for this you would need to have the fix of the related bug reported also by you (https://www.chemaxon.com/forum/viewtopic.php?p=57036#57036 ).
A workaround would be to ignore the charges with "charge:i" option. This would work for this case, but probably would generate false positive hits, depending on your use cases. Would this work for you?
Best regards,
Norbert
User 81a38f9467
24-06-2014 00:00:20
Hi Norbert,
> Would this work for you?
No.
We do t:d searches during registration to avoid the registration of duplicates. The charge:i option doesn't work with a t:d search, though I realize that 't:f charge:i' would function OK. I'll wait for a fix rather than change the query to something other than a duplicate search.
Tim
ChemAxon d4fff15f08
25-06-2014 13:36:13
Hi Tim,
Clear!
The corrected Mesomerize standardization action will be available in the upcoming 6.4 version.
Norbert.
ChemAxon d4fff15f08
09-07-2014 10:51:59
Hi Tim,
The new version (14.7.7.0) of JChem Base is available for download. It contains the fix for the bug reported by you.
Best regards,
Norbert