Ring perception

A
colleague asked me to relay to you some apparent cases where the
calculation of how many rings an atom is in some way flawed. I
enclose the SMILES strings I was sent, and the explanation. It seems to be where rings appear to be contained within rings. Is this working as intended, or is there some obscure issue.

      


CC1(C)[C@H]2C[C@@H]1[C@]1(CO1)CC2>>CC1(C)[C@H]2C[C@@H]1[C@@H](CC2)C=O

      


      The reactant in this reaction apparently has issues in Marvin

      calculation, the carbon in the epoxide ring is part of 2 other

      rings, but is only reported as being in 2 in total (not 3)

      


      


O=C1[C@H]2C[C@@H]3C[C@H](C[C@H]1C3)C2.CO>>C1O[C@]11[C@H]2C[C@@H]3C[C@H](C[C@H]1C3)C2

      


      There are multiple issues here with the ring perception, one side

      versus the other, not really sure how to explain it, but it seems

      that it's around the epoxide again in the product.

     

Dear James,

I can not give an exact reply because I do not know how you obtain the information about the number of rings in the molecule. Do you use our basic structural API? Which method? I can only guess that the number of rings is defined by the Smallest Set of Smallest Rings (SSSR), which is usually the case in our ring calculation. In this case rings which are the combination of smaller rings are not included in the contained set of rings. This means that the given answer is correct, for each atom in your molecule there are maximum two rings (not covered by other rings) which contain it.

Best regards,

Zsuzsa

Yes, we use SSSR for this calculation (MoleculeGraph.getSSSRIdxesForAtoms()[atomindex]), so that would explain that. Out of interest is the total number of rings available through the API at all? It would be nice to get a comprehensive value as well.

Dear James,

The method MoleculeGraph.getAromaticAndAliphaticRings() returns all aromatic and all aliphatic rings separately, for details please refer to our API documentation at http://www.chemaxon.com/jchem/doc/dev/java/api/ .

Regards,

Zsuzsa