Molecule Diameter

[TUstudents]

For a ML model, I am looking for a descriptor that correlates with the molecule size and is calculate-able given a SMILES string. I found that the 'TopologicalIndex" gives good results, but the TopologicalIndex.Diameter is not useful for ionic bound crystals (1E8). Given only the SMILES string, the GeometricalIndex.Diameter is not calculate-able because of the missing 3d info. Descriptors that calculate the VDW Volume struggle on metal ions.

Which other descriptor would you suggest?

import pandas as pd
from rdkit import Chem
from mordred import Weight, TopologicalIndex, 	GeometricalIndex, VdwVolumeABC, Calculator, is_missing, descriptors

compound = [['Citrate', '126-44-3','C(C(=O)[O-])C(CC(=O)[O-])(C(=O)[O-])O'],
    ['Citric acid', '77-92-9', 'C(C(=O)O)C(CC(=O)O)(C(=O)O)O'],
    ['Sodium citrate','68-04-2','C(C(=O)[O-])C(CC(=O)[O-])(C(=O)[O-])O.[Na+].[Na+].[Na+]'],
    ['Copper(II) citrate','3164-31-6','C(C(=O)[O-])C(CC(=O)[O-])(C(=O)[O-])O.C(C(=O)[O-])C(CC(=O)[O-])(C(=O)[O-])O.[Cu+2].[Cu+2].[Cu+2]'],
    ['Alverine Citrate', '5560-59-8', 'CCN(CCCC1=CC=CC=C1)CCCC2=CC=CC=C2.C(C(=O)O)C(CC(=O)O)(C(=O)O)O'],
    ['Sildenafil Citrate', '171599-83-0', 'CCCC1=NN(C2=C1N=C(NC2=O)C3=C(C=CC(=C3)S(=O)(=O)N4CCN(CC4)C)OCC)C.C(C(=O)O)C(CC(=O)O)(C(=O)O)O']]


df = pd.DataFrame(compound, columns=['Name', 'CAS', 'SMILES'])
calc = Calculator()
calc.register(Weight.Weight(True,True))
calc.register(TopologicalIndex.Diameter)
calc.register(GeometricalIndex.Diameter3D)
calc.register(VdwVolumeABC.VdwVolumeABC())

mols = [Chem.MolFromSmiles(str(x)) for x in df['SMILES'].values]
size=calc.pandas(mols)

#	aMW	Diameter	GeomDiameter	Vabc
0	189.099	6	missing 3D coordinate (GeomDiameter/mordred._g...	161.999894
1	192.123	6	missing 3D coordinate (GeomDiameter/mordred._g...	165.954582
2	258.069	100000000	missing 3D coordinate (GeomDiameter/mordred._g...	unknown atom type (Vabc/nARing/mordred.RingCou...
3	568.836	100000000	missing 3D coordinate (GeomDiameter/mordred._g...	unknown atom type (Vabc/nARing/mordred.RingCou...
4	473.566	100000000	missing 3D coordinate (GeomDiameter/mordred._g...	469.095835
5	666.710	100000000	missing 3D coordinate (GeomDiameter/mordred._g...	574.092163