structure to numbering rather than name ?

Is there a way to get from a structure (drawn or SMILES) to a IUPAC numbering of all the carbon atoms for a compound? The result needs to give the same results regardless of how the structure is drawn (so independent of the order used to draw the parts of the structure) or the choice of SMILES.

We could for instance generate the following numbering, represented in cxsmiles:

OC(=O)c1ccccc1CCC1CCCC1 |$_AV:;;;1;6;5;4;3;2;1;2;1;5;4;3;2$|

(see attached image below).

The numbering is consistent with the IUPAC rules, and is independent of the smiles representation or order of the atoms and bonds.

How would you use this feature? Would you need batch conversion (command line or API), or rather interactive display, for instance in Marvin Sketch?

You will neeed a Structure to Name license for this feature. Do you have one, or would you like to get an evaluation license and discuss conditions?

Thanks for the reply. However this is not what I was having in mind. The numbering would have to be unique for each C atom and, preferably, without unexpected jumps between numbers for connected atoms. The atom numbering that Marvin Sketch shows now is mostly suitable if it could be restricted to the C atoms although it sometimes will number an atom out of order.

I see. If you want a unique number for each carbon atom, it cannot be based on the IUPAC numbering, since that assigns numbers locally to each fragment (ring system and chain). Your request could be doable with a custom modification of the numbering algorithm. I moved the topic, it will be answered by the appropriate team.

Thanks for the information and the move to a new thread.

How would you like to use this feature?

Is an API-based solution fine? Or would you rather do it in MarvinSketch?

With API code it can be solved quite easily, I can provide a sample code if needed.

For me (us) the ideal would be a function in molconvert that would give a labelled graphics image. Somthing like

molconvert smiles:numC molecule.smiles -o molecule.jpg

where numC tells molconvert to just number the C atoms in as a connected way as possible. The numbers should be monotonically increasing with as few discontinuities as can be made. The resulting scheme would be used to label X-ray structures as well as for identifying (labelling) chiral carbon centers.

The second choice would be a choice inside Marvin Sketch

Thank you for the clarification. Currently we don't have a solution for this.

We will implement this new feature as soon as possible.

Thanks for considering adding this functionality.