smiles generation

Greetings, I have a question regarding how (unique) SMILES are generated. Basically I have two instances of Molecule's and the output of toFormat("smiles:u") yield the following:





CS(=O)(=O)Nc1cccc(c1)-c2nc(-NCc3ccccc3)c4ccccc4n2


CS(=O)(=O)Nc1cccc(c1)-c2nc(NCc3ccccc3)c4ccccc4n2





Can you tell me what's the difference between the two instances? More generally, are there specific rules for which explicit single bonds are generated instead of implicit ones?





Thanks for any pointer,


Trung





ps: I'm using JChem 3.2.12

I moved this topic to " Structure editing, viewing and file formats" forum. My colleagues will answer you soon.





Zsolt

Hi,





I guess the two molecules just looks identical but they are not.


Could you please attach the molecules in original format so I can point out the difference.





Thank you


Andras

These two molecules are just different tautomer forms. After some debugging, I found that the second explicit single bond in the following





CS(=O)(=O)Nc1cccc(c1)-c2nc(-NCc3ccccc3)c4ccccc4n2





was due to invalid stereo annotation. A mistake on my part for using MolBond.setFlags() instead of MolBond.setType(). With this fix, I got both instances to generate identical SMILES:





CS(=O)(=O)Nc1cccc(c1)-c2nc(NCc3ccccc3)c4ccccc4n2


CS(=O)(=O)Nc1cccc(c1)-c2nc(NCc3ccccc3)c4ccccc4n2





For the purpose of standardization/canonicalization, this is good enough for me, though I'm still curious as to the general condition for which explicit single bonds are generated.





Trung

The default bond type between to aliphatic atoms is single bond, the default bond type between to aromatic atoms is aromatic bond.





This is why the single bond must be explicitly specified between two aromatic atoms of biphenyl type systems.

Exactly

Ok thanks. I don't recall reading about this in daylight documentation on SMILES. So according this rule, by removing the explicit single bond as in





CS(=O)(=O)Nc1cccc(c1)-c2nc(NCc3ccccc3)c4ccccc4n2


CS(=O)(=O)Nc1cccc(c1)c2nc(NCc3ccccc3)c4ccccc4n2





the single bond connecting two aromatic rings is now an aromatic bond? Clearly this is not the case here. Is this just a convention (perhaps for readability) or is there a deeper reason behind it?





Thanks,


Trung

As you have checked, we import the following SMILES


CS(=O)(=O)Nc1cccc(c1)c2nc(NCc3ccccc3)c4ccccc4n2


with single bonds between the aromatic rings.





Our philosophy is the following:


The single bond is written explicitly if it is between two aromatic atoms.


This helps readability and also ensures that other (third party) SMILES importers are getting the same structure as the exported.





Andras