About planarity of amides and other planar groups in RDKit generated 3D structures

[gcincilla]

Dear RDKit lovers,

Very recently I tested the new option added in the RDKit Add Conformers node called “use experimental torsion angle terms”. As highlighted in this thread, this option decreases the likelihood to obtain a structure with cis amide groups. This new option works very well and it avoids the generation of such problematic structures (at least in one of my test sets). In my opinion this makes RDKit the best option for 3D structure generation in KNIME, so thanks a lot to all the developers for this great improvement!

Nevertheless I think there is a couple of more issues in the generation of 3D structures that need to be addressed and for which I would like to have your opinion:

1. The first is related to the planarity of amide groups. In several cases I obtain molecules with non-planar amides. The dihedral of such groups can reach values up to around 20º for secondary amides and up to around 40º for tertiary amides. Can the generation of such structures be avoided? It would be great if the planarity of amides could somehow be promoted in future RDKit releases.

2. The 3D coordinates generator produces non planar sulfonamide nitrogens, even in case of primary sulfonamides. Shouldn't these groups be planar?

I will appreciate any comment regarding these issues.

Gio

[greglandrum]

Hi Gio,

Thanks for pointing out the problems. Assuming that we can reproduce the bad geometries, we will look into fixing them. It would really help if you could provide the feedback about what you're seeing, ideally with example structures, in the RDKit github issue tracker here: https://github.com/rdkit/rdkit/issues

Best,

-greg

[gcincilla]

Hi Greg,

Thanks a lot for your quick reply. I open this issue to the RDKit github issue tracker, as you suggested. For sake of completeness I upload here a workflow generating some example structures.

• rdkit_3d_generated_problematic_structures.knwf