Question about tautomer comparison

Hi,

 

Previously, you clarified that tautomer comparisons by API and/or cartridge are based on generic tautomer for each structure.

 

For example, these two SMILES are considered tautomers because they produce same generic tautomer:

 

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

OC(O)[C@H](O)[C@@H](O)C(O)=O

 

However, some of our chemists would prefer that these two SMILES are not identified as tautomers because there is a much higher activation energy between these two structures than, for example, a typical keto/enol tautomer pair .. and so they say that these two structures will not readily inter-covert under typical conditions and so they should be considered distinct.

 

Is there some API/cartridge method/suggestion for a tautomer comparison which considers keto/enol tautomers as a match, but would not consider the above set of SMILES to be tautomers?

 

Thank you for any suggestions.

 

Best regards,

 

Andrew

Dear Andrew,

 

The given two structures differ in a double bond, one is C4H6O6 while the other is C4H8O6. None of their tautomers would match in a duplicate search.

 

However, generally speaking about tautomer searches, the duplicate, full fragment, and full searches use generic tautomers to compare the two (query and target) structures. During searches, unfortunately, there is not possible yet to have structures excluded from a general search option (i.e tautomerSearch:y/n) based on other considerations (e.g. pH, activation energy, excluded list, etc.). A possible workaround would be to play with the (tetrahedral) stereo information and with the ignreStereo related search options accordingly. You may find a description about these search options here: http://www.chemaxon.com/jchem/doc/user/Jcsearch.html

Even so, this would not be applicable for any tautomer pairs, just for some of them.

 

Hope I could give you some ideas.

 

Best regards,

Norbert

Hi,

Thank you for the information.

I misreported one of the SMILES in the original post .. the two SMILES with undesirable tautomer match are:

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

O[C@H]([C@@H](O)C(O)=O)C(O)=O

I understand this does not change your reply, but I include these here to correct my error .. smile.

 

Best,

Andrew

Hi Andrew,

 

Thank you for clarifying the problematic structures, it was obvious that it was just an editing error.

In the light of this, my reply would change a bit, now, as a chemist I totally understand your concern about tautomer recognition. Irrespectively of it, generic tautomer regions for both structures cover the whole molecule (the single regions inside the molecule are quasi overlapping) and as a consequence, the tautomer search will correctly (algorithmically) identify them tautomers.

Thank you for dragging our attention to this special behaviour of generic tautomer generation. We will consider to handle these special cases differently.

 

Best regards,

Norbert

Thank you for the response .. I hope you can come up with a solution for this soon .. smile.

Best,

Andrew