Unexpected apparition of RAD

[mpagni12]

Describe the bug

RAD may appears ex nihilo in MOL block when an heavy atom is present.

To Reproduce

from rdkit import Chem

# The first mol block is KEGG C06696 for metalic lead
mol_block_1 = """


  1  0  0  0  0  0  0  0  0  0999 V2000
   37.3800  -22.7500    0.0000 Pb  0  0  0  0  0  0  0  0  0  0  0  0
M  END
"""
mol_block_2 = """
     RDKit          2D

  1  0  0  0  0  0  0  0  0  0999 V2000
   37.3800  -22.7500    0.0000 Pb  0  0  0  0  0  2  0  0  0  0  0  0
M  END
""" 
mol_block_3 = """
     RDKit          2D

  1  0  0  0  0  0  0  0  0  0999 V2000
   37.3800  -22.7500    0.0000 Pb  0  0  0  0  0  2  0  0  0  0  0  0
M  RAD  1   1   3
M  END
"""
print( Chem.MolToMolBlock( Chem.MolFromMolBlock( mol_block_1 )) == mol_block_2 )
print( Chem.MolToMolBlock( Chem.MolFromMolBlock( mol_block_2 )) == mol_block_3 )

yields

True
True

Expected behavior

True
False

Configuration

• RDKit version: 2022.9.1
• OS: macOS Monterey
• Python version (if relevant): Python 3.10.8
• Are you using conda? no
• If you are not using conda: how did you install the RDKit? pip3 install rdkit

Additional context There are many other entries in KEGG with an heavy atom that display the same problem. The one I am reporting here is pretty "compact".

Thank you for maintaining RDKit,

Marco

[rapodaca]

mol_block_2 and mol_block_3 are tickling the atomic valence (vvv) attribute. Setting it to anything other than 0 sets the valence to that value, and turns off implicit hydrogen detection, which is what ordinarily sets valence.

Under the pre-2014 regime, Pb was considered one of those atoms whose hydrogen count could be computed ("implicit"). After, this was disabled and it was no longer possible to compute implicit H on Pb and other elements ("MDL Valence-Mageddon").

Under pre-2014:

• mol_block_1 encodes PbH2 (default valence 2, w/ implicit H detection).
• mol_block_2 encodes PbH2 (explicit valence 2)
• mol_block_3 encodes PbH2 (explicit valence 2 w/ 2e radical)

Under post-2014:

• mol_block_1 encodes Pb (no implicit H)
• mol_block_2 encodes PbH2 (valence 2 and so 2 implicit H)
• mol_block_3 encodes PbH2 (valence to, 2 implicit H, and 2e electron)

It's possible that the results reflect a desire to eliminate valence ambiguity by explicitly setting valence attribute. And then RDKit is noticing that valence has been set and so sets the radical attribute.

[greglandrum]

I believe that this is a bug (potentially more than one bug), but this stuff is a pain, so I'm going to have to spend some time digging to be sure.

[github-actions[bot]]

This issue was marked as stale because it has been open for 90 days with no activity.

[github-actions[bot]]

This issue was closed because it has been inactive for 14 days since being marked as stale.