RDKit stores absolute configuration around atoms using parity (which is based on atom index), not CIP codes. This has oodles of benefits, for instance you can represent absolute configuration of atoms in achiral structures like 1,4-dimethylcyclohexane. So, yeah - AssignStereochemistry is intended to produce different results with different atom sequences.
If you need Cahn/Ingold/Prelog-style R or S for display, maybe try: from rdkit import Chem print(Chem.FindMolChiralCenters(mol)) https://github.com/rdkit/rdkit/blob/2d8bb6c1687a2fa78f66fc61966544e5f44f157c/rdkit/Chem/__init__.py#L80 *dan nealschneider* | senior developer [image: Schrodinger Logo] <https://www.schrodinger.com/> On Wed, Sep 25, 2019 at 2:35 PM Zoltan Takacs <zozo...@gmail.com> wrote: > Dear RDkitters, > > I am playing around with the assignstereochemistry function and I am > getting confused. If I change the numbering of the atoms but not their > coordinates (for example switch to canonical numbering) I get a different > answer. In fact it only seems to work right when I switch to canonical > numbering with openbabel. > > What rules does the function follow? Is there any role played by the > numbering of atoms? Do they have to be numbered in any special way for > example the groups need to follow each other when assigning CIP priority? > > Thanks > > _______________________________________________ > Rdkit-discuss mailing list > Rdkit-discuss@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/rdkit-discuss >
_______________________________________________ Rdkit-discuss mailing list Rdkit-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
