Substructure search confusion

Hi,

I am trying to find the following structure in my database by performing a substructure search:

CCN(C(=O)C1=C(O)C2=C(Cl)C=CC=C2N(C)C1=O)C1=CC=CC=C1

If I perform my substructure search with this exact SMILES, I retrieve this compound as expected. If I remove the blue highlighted '=O', I still retrieve this compound. However if I keep the blue one and remove the red highlighted '=O' instead, I no longer retrieve this compound.

Am I doing something wrong here? The red '=O' extends from a ring, whereas the blue one does not. Do I need to add in some specific SMILES features to make this search work? I get the same strange behavior if I use lower case c characters.

Thank you for any clarification.

Hi,

I suggest you to use the Marvin Sketch desktop GUI to conveniently visualize SMILES, edit structures and obtain the modified SMILES (Edit -> Source)

During search JChem converts single and double bonds of aromatic rings in both query and target to aromatic bond type, then an aromatic bond will only match an aromatic bond during graph search.

Please find more details here:

https://docs.chemaxon.com/display/jchembase/Standardization

The two modified structures are the following:

CCN(CC1=C(O)C2=C(Cl)C=CC=C2N(C)C1=O)C1=CC=CC=C1

CCN(C(=O)C1=C(O)C2=C(Cl)C=CC=C2N(C)C1)C1=CC=CC=C1 

If you paste the original structure, or the first modified structure into Marvin Sketch and aromatize them, you can see all rings are aromatic.

(Structure -> Aromatic Form -> Convert to Aromatic form, the default method is "general") 

For the second modified structure the ring from which the oxygen was removed will no longer considered as aromatic, therefore there will be no match.

Please also see the Vague Bond level search options, where depending on the desired application bond matching rules can be loosened up in stages, even up to totally ignoring bond types:

https://docs.chemaxon.com/pages/viewpage.action?pageId=22217121

Thank you very much for the explanation. Indeed the issue is that my search query in the second case contains a non-aromatic ring, and will not match to my aromatic target. If I aromatize the ring in question, the search will retrieve the target parent structure. Thank you!

FYI, the URL you provided contains an additional character at the end which breaks the link.

I'm glad this helped.

I have fixed the link above for future visitors of the topic, thank you for pointing it out.

Szilard