Bug in substructure search with List or A(ny) atom in query

User f67d4188b6

10-10-2012 15:16:14

2 molecules:


C=C1NC(=O)C(=N)S1
C=C1NC(=O)C(=C)S1


2 queries:


C=C1NC(=O)C(=*)S1
C=C1NC(=O)C(=[C,N])S1


 


In IJC 5.11.1 we get only the molecule with the Nitrogen with both queries, but expect to hit both.

ChemAxon 2bdd02d1e5

11-10-2012 13:28:43

Thank you for your report. We are looking at this issue.

ChemAxon abe887c64e

16-10-2012 09:36:08

Hi,


This is known issue based on the 'general' type aromatization we use by default. http://www.chemaxon.com/marvin/help/sci/aromatization-doc.html#basic


We would like to suggest that you apply one of the following workarounds:



Please let us know if you have any further questions.


Krisztina

User f67d4188b6

17-10-2012 09:14:04

Thank you for this.


We are using InstantJChem with a mysql database, the standardization step is allready done automatically, including the Aromatize step.
Changing the Vague Bond setting in the search options does not help, and it is on level2 by default.


The one that did work was to use "single or aromatic" bond type in the ring.
We noticed that you can check the "Display as standardized" box in the Visual properties of a structure column, and only then is the difference visible between the 2 groups.

ChemAxon abe887c64e

17-10-2012 09:37:59










Ellert van Koperen wrote:

Thank you for this.


We are using InstantJChem with a mysql database, the standardization step is allready done automatically, including the Aromatize step.
Changing the Vague Bond setting in the search options does not help, and it is on level2 by default.


The one that did work was to use "single or aromatic" bond type in the ring.
We noticed that you can check the "Display as standardized" box in the Visual properties of a structure column, and only then is the difference visible between the 2 groups.



I'm happy you have found a working solution. However, in my tests with IJC 5.11.1 the Vague Bond setting: Ring bonds or aromatic (this is level 2) worked as well. The default setting is level 1.

User f67d4188b6

23-10-2012 11:50:39

You are correct, i started to count from 1 and should have started counting from 0 in the options.
With the option set to "single or aromatic" bond type in the ring IJC does find them all.


 


However, we found that this same issue occurs in the clean-2d filter in the Standardizer (Knime node)
We intended to use basically the same query to set the orientation of all molecules in a list using the template option in clean2d.
It only worked if we defined the bonds in the 5-membered ring as of 'any' type.

ChemAxon a9ded07333

31-10-2012 10:36:24

Hi Ellert,


Unfortunately you are right, you should change the query when using Standardizer in Knime.
You may try to change the bonds to 'single or aromatic' or 'double or aromatic' instead of 'any' - that can improve efficiency.


Best regards,
Tamás