[Rdkit-discuss] cis/trans info lost when generating SMILES from molfile

[Eloy F. <elo...@gm...>]

Hi RDKitters,

I'm having trouble writing SMILES including cis/trans info for some
molecules I load from molfile using rdkit. Openbabel and indigo are
generating the expected SMILES.

https://www.ebi.ac.uk/chembl/compound/inspect/CHEMBL501674

*# RDKIT*
mol = Chem.MolFromMolFile('CHEMBL501674.mol')
Chem.MolToSmiles(mol, isomericSmiles=True)
'CO[C@H]1[C@H](OC[C@H]2[C@@H]3O[C@@H]3C=CC(=O)[C@H](C)CC[C@H](O[C@@H]3O[C@H
](C)C[C@H](O)[C@H]3O)[C@@H](C)C=CC(=O)O[C@@H]2C)O[C@H](C)[C@@H](O)[C@H]1OC'

*# OPENBABEL*
mol_ob = pybel.readfile('mol', 'CHEMBL501674.mol')
mol_smiles.write('can')
'CO[C@H]1[C@H](OC[C@@H]2[C@@H](C)OC(=O)*/C=C/*[C@H](C)[C@H](CC[C@H](C(=O)
*/C=C/*[C@@H]3[C@H]2O3)C)O[C@@H]2O[C@H](C)C[C@@H]([C@H]2O)O)O[C@@H]([C@H
]([C@H]1OC)O)C'

*# INDIGO*
mol = indigoObject.loadMoleculeFromFile('CHEMBL501674.mol')
mol.smiles()
'C1(=O)[C@H](C)CC[C@H](O[C@@]2([H])[C@H](O)[C@@H](O)C[C@@H](C)O2)[C@
@H](C)C=CC(=O)O[C@H](C)[C@@H](CO[C@]2([H])O[C@H](C)[C@@H](O)[C@@H](OC)[C@H
]2OC)[C@]2([H])O[C@]2([H])C=C1 |*t:21,51*
,&1:2,6,8,10,12,15,18,25,27,30,33,35,37,40,43,46|'

Indigo is using chemaxon extended notation, but is also . recognising
t:21,51.


If I check double bond stereo info for the molecule:

for bond in mol.GetBonds():
    if bond.GetBondType() == Chem.BondType.DOUBLE:
        print(bond.GetBeginAtomIdx(), bond.GetEndAtomIdx(),
bond.GetStereo())
6 7 STEREONONE
11 12 STEREONONE
0 17 STEREONONE
5 43 STEREONONE

No E/Z info in bonds.

inchi bond stereo layer generated from the molfile: /b12-10+,13-9+
from rdkit generated smiles(with and without Compute2DCoords):
/b12-10-,13-9-
from obabel generated smiles: /b12-10+,13-9+

This might be a bug in the piece of code that detects bond stereo info from
the molfile or maybe... I'm just missing something :P

Thanks for your great job!