How to handle unacceptable SMILEs?

[RunshengSong]

Hi, I used your package to calculate descriptors for thousands chemicals all together. My code:

mols = [Chem.MolFromSmiles(smi) for smi in all_SMILEs]
df = calc.pandas(mols)

where "all_SMILEs" contain thousands SMILEs, but some of them are not acceptable by mordred (or rdkit ?). For example, this one: O=C(c1cc(Cl)c(Cl)n1)c2c(O)c(cc(Cl)c2O)CCCCCC

And it throw this error:

---------------------------------------------------------------------------
ArgumentError                             Traceback (most recent call last)
<ipython-input-15-78193e400eba> in <module>()
     14     mols = [Chem.MolFromSmiles(smi)]
     15 
---> 16     df = calc.pandas(mols)
     17 
     18 mols = [Chem.MolFromSmiles(smi) for smi in all_SMILEs]

/home/runsheng/anaconda2/envs/CLiCC/lib/python2.7/site-packages/mordred/_base/calculator.pyc in pandas(self, mols, nproc, nmols, quiet, ipynb, id)
    293         return pandas.DataFrame(
    294             self.map(mols, nproc, nmols, quiet, ipynb, id),
--> 295             columns=[str(d) for d in self.descriptors]
    296         )
    297 

/home/runsheng/anaconda2/envs/CLiCC/lib/python2.7/site-packages/pandas/core/frame.pyc in __init__(self, data, index, columns, dtype, copy)
    298         elif isinstance(data, (list, types.GeneratorType)):
    299             if isinstance(data, types.GeneratorType):
--> 300                 data = list(data)
    301             if len(data) > 0:
    302                 if is_list_like(data[0]) and getattr(data[0], 'ndim', 1) == 1:

/home/runsheng/anaconda2/envs/CLiCC/lib/python2.7/site-packages/mordred/_base/parallel.pyc in parallel(self, mols, nproc, nmols, quiet, ipynb, id)
     41     try:
     42         with self._progress(quiet, nmols, ipynb) as bar:
---> 43             for result in [do_task(m) for m in mols]:
     44                 r, err = get_result(result)
     45 

/home/runsheng/anaconda2/envs/CLiCC/lib/python2.7/site-packages/mordred/_base/parallel.pyc in do_task(mol)
     36 
     37     def do_task(mol):
---> 38         args = Context.from_calculator(self, mol, id)
     39         return pool.apply_async(worker, (args,))
     40 

/home/runsheng/anaconda2/envs/CLiCC/lib/python2.7/site-packages/mordred/_base/context.pyc in from_calculator(cls, calc, mol, id)
     54     @classmethod
     55     def from_calculator(cls, calc, mol, id):
---> 56         return cls.from_query(mol, calc._require_3D, calc._explicit_hydrogens, calc._kekulizes, id)
     57 
     58     def get_coord(self, desc):

/home/runsheng/anaconda2/envs/CLiCC/lib/python2.7/site-packages/mordred/_base/context.pyc in from_query(cls, mol, require_3D, explicit_hydrogens, kekulizes, id)
     24     @classmethod
     25     def from_query(cls, mol, require_3D, explicit_hydrogens, kekulizes, id):
---> 26         n_frags = len(Chem.GetMolFrags(mol))
     27 
     28         if mol.HasProp('_Name'):

ArgumentError: Python argument types in
    rdkit.Chem.rdmolops.GetMolFrags(NoneType)
did not match C++ signature:
    GetMolFrags(RDKit::ROMol mol, bool asMols=False, bool sanitizeFrags=True)

How should I handle SMILEs that is not acceptable? This error message would be more useful if it tells me which SMILEs is not acceptable so I can remove it myself.

Thanks.

[philopon]

I'm sorry for the late reply.

It seems not acceptable SMILES by rdkit (not mordred). Chem.MolFromSmiles returns None if it is given invalid SMILES. You can handle error by checking None.

example:

mols = []
for smi in all_SMILES:
    mol = Chem.MolFromSmiles(smi)
    if mol is None:
        # handing error
        print("cannot parse SMILES: {}".format(smi))
        continue

    mols.append(mol)

I think this error message it is little unkind. I think to fix it.

Thanks.