Zero & coordinate bonds are being taken into account for chirality

[ricrogz]

I'm seeing the following using current master:

Consider this input mol. It has to be a mol block because of the zero bonds, which are not supported on SMILES. They are actually non standard on mol blocks too!

from rdkit import Chem

m = Chem.MolFromMolBlock("""
     RDKit          3D

  0  0  0  0  0  0  0  0  0  0999 V3000
M  V30 BEGIN CTAB
M  V30 COUNTS 15 18 0 0 0
M  V30 BEGIN ATOM
M  V30 1 C -0.136359 0.025241 -0.986870 0
M  V30 2 C 0.211183 -0.810922 0.138318 0
M  V30 3 C -0.446638 -0.713741 1.305561 0
M  V30 4 C -1.141107 0.914647 -0.916429 0
M  V30 5 R -1.628248 -0.983190 -0.411960 0
M  V30 6 H 0.392055 -0.106505 -1.920607 0
M  V30 7 H 0.974038 -1.568492 0.017171 0
M  V30 8 H -0.209921 -1.406535 2.084966 0
M  V30 9 H -1.378909 1.482059 -1.807349 0
M  V30 10 C -1.544607 0.306162 1.588191 0
M  V30 11 C -1.946856 1.186683 0.358271 0
M  V30 12 H -1.207983 0.944410 2.407927 0
M  V30 13 H -2.419549 -0.225146 1.965589 0
M  V30 14 H -3.006492 1.040978 0.144313 0
M  V30 15 H -1.830875 2.240146 0.620809 0
M  V30 END ATOM
M  V30 BEGIN BOND
M  V30 1 1 2 1
M  V30 2 2 3 2
M  V30 3 2 4 1
M  V30 4 0 5 1
M  V30 5 0 5 2
M  V30 6 0 5 3
M  V30 7 0 5 4
M  V30 8 1 1 6
M  V30 9 1 2 7
M  V30 10 1 3 8
M  V30 11 1 4 9
M  V30 12 1 10 3
M  V30 13 1 11 4
M  V30 14 1 11 10
M  V30 15 1 12 10
M  V30 16 1 13 10
M  V30 17 1 14 11
M  V30 18 1 15 11
M  V30 END BOND
M  V30 END CTAB
M  END
""")

----
[12:13:04] bond with order 0 found on line 28. This is not part of the MDL specification.
[12:13:04] bond with order 0 found on line 29. This is not part of the MDL specification.
[12:13:04] bond with order 0 found on line 30. This is not part of the MDL specification.
[12:13:04] bond with order 0 found on line 31. This is not part of the MDL specification.

Up to this point, things are fine:

m.Debug()                                                                                                                                           
Atoms:
	0 6 C chg: 0  deg: 3 exp: 3 imp: 1 hyb: 4 arom?: 0 chi: 0
	1 6 C chg: 0  deg: 3 exp: 3 imp: 1 hyb: 4 arom?: 0 chi: 0
	2 6 C chg: 0  deg: 3 exp: 3 imp: 1 hyb: 4 arom?: 0 chi: 0
	3 6 C chg: 0  deg: 3 exp: 3 imp: 1 hyb: 4 arom?: 0 chi: 0
	4 0 * chg: 0  deg: 4 exp: 0 imp: 0 hyb: 0 arom?: 0 chi: 0
	5 6 C chg: 0  deg: 2 exp: 2 imp: 2 hyb: 4 arom?: 0 chi: 0
	6 6 C chg: 0  deg: 2 exp: 2 imp: 2 hyb: 4 arom?: 0 chi: 0
Bonds:
	0 1->0 order: 1 conj?: 0 aromatic?: 0
	1 2->1 order: 2 conj?: 0 aromatic?: 0
	2 3->0 order: 2 conj?: 0 aromatic?: 0
	3 4->0 order: 0 conj?: 0 aromatic?: 0
	4 4->1 order: 0 conj?: 0 aromatic?: 0
	5 4->2 order: 0 conj?: 0 aromatic?: 0
	6 4->3 order: 0 conj?: 0 aromatic?: 0
	7 5->2 order: 1 conj?: 0 aromatic?: 0
	8 6->3 order: 1 conj?: 0 aromatic?: 0
	9 6->5 order: 1 conj?: 0 aromatic?: 0

But if we assign stereo, we start seeing weird things:

Chem.AssignStereochemistryFrom3D(m)                                                                                                                 
m.Debug()                                                                                                                                           
Atoms:
	0 6 C chg: 0  deg: 3 exp: 3 imp: 1 hyb: 4 arom?: 0 chi: 2
	1 6 C chg: 0  deg: 3 exp: 3 imp: 1 hyb: 4 arom?: 0 chi: 1
	2 6 C chg: 0  deg: 3 exp: 3 imp: 1 hyb: 4 arom?: 0 chi: 2
	3 6 C chg: 0  deg: 3 exp: 3 imp: 1 hyb: 4 arom?: 0 chi: 1
	4 0 * chg: 0  deg: 4 exp: 0 imp: 0 hyb: 0 arom?: 0 chi: 1
	5 6 C chg: 0  deg: 2 exp: 2 imp: 2 hyb: 4 arom?: 0 chi: 0
	6 6 C chg: 0  deg: 2 exp: 2 imp: 2 hyb: 4 arom?: 0 chi: 0
Bonds:
	0 1->0 order: 1 conj?: 0 aromatic?: 0
	1 2->1 order: 2 conj?: 0 aromatic?: 0
	2 3->0 order: 2 conj?: 0 aromatic?: 0
	3 4->0 order: 0 conj?: 0 aromatic?: 0
	4 4->1 order: 0 conj?: 0 aromatic?: 0
	5 4->2 order: 0 conj?: 0 aromatic?: 0
	6 4->3 order: 0 conj?: 0 aromatic?: 0
	7 5->2 order: 1 conj?: 0 aromatic?: 0
	8 6->3 order: 1 conj?: 0 aromatic?: 0
	9 6->5 order: 1 conj?: 0 aromatic?: 0

RDkit found 5 chiral atoms: 4 C atoms with 1 implicit H + 1 single + 1 double + 1 zero bonds, and the dummy atom, with 4 zero bonds (!!!!).

And here is where things explode:

Chem.MolToMolBlock(m)                                                                                                                               
[12:19:02] 

****
Invariant Violation
no eligible neighbors for chiral center
Violation occurred on line 240 in file /rdkit_builder/rdkit/Code/GraphMol/FileParsers/MolFileStereochem.cpp
Failed Expression: nbrScores.size()
****

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
Input In [6], in <cell line: 1>()
----> 1 Chem.MolToMolBlock(m)

RuntimeError: Invariant Violation
	no eligible neighbors for chiral center
	Violation occurred on line 240 in file Code/GraphMol/FileParsers/MolFileStereochem.cpp
	Failed Expression: nbrScores.size()
	RDKIT: 2022.09.1pre
	BOOST: 1_76

This is happening inside WedgeMolBonds(), when we try to find a bond to wedge for the dummy atom: as it only has zero bonds, of course, we can't find a single bond we can wedge to encode the parity in the mol block.

[lonngxiang]

The same error, how to solve?

[greglandrum]

@lonngxiang : we can't help if you don't provide the sample input which caused the problem you are seeing.