def calc_molfeatures(mol: Chem.Mol, features_generator: FeaturesGenerator):
if mol is not None and mol.GetNumHeavyAtoms() > 0:
molfeatures = features_generator(mol)
# for H2
elif mol is not None and mol.GetNumHeavyAtoms() == 0:
# not all molfeatures are equally long, so use methane as dummy
# molecule to determine length
molfeatures = np.zeros(len(features_generator(
Chem.MolFromSmiles('C'))))
else:
molfeatures = None
return np.asarray(molfeatures)
def CalculateFlexibility(mol: Chem.Mol) -> float:
"""Calculate Kier molecular flexibility index."""
kappa1 = CalculateKappaAlapha1(mol)
kappa2 = CalculateKappaAlapha2(mol)
A = mol.GetNumHeavyAtoms()
phi = kappa1 * kappa2 / (A + 0.0)
return round(phi, 3)
def CalculateKappa1(mol: Chem.Mol) -> float:
"""Calculate molecular shape index for one bonded fragment."""
P1 = mol.GetNumBonds(onlyHeavy=1)
A = mol.GetNumHeavyAtoms()
denom = P1 + 0.0
if denom:
kappa = (A) * (A - 1)**2 / denom**2
else:
kappa = 0.0
return round(kappa, 3)
def CalculateKappa2(mol: Chem.Mol) -> float:
"""Calculate molecular shape index for two bonded fragments."""
P2 = len(Chem.FindAllPathsOfLengthN(mol, 2))
A = mol.GetNumHeavyAtoms()
denom = P2 + 0.0
if denom:
kappa = (A - 1) * (A - 2)**2 / denom**2
else:
kappa = 0.0
return round(kappa, 3)
def CalculateKappaAlapha1(mol: Chem.Mol) -> float:
"""Calculate molecular shape index for one bonded fragment."""
P1 = mol.GetNumBonds(onlyHeavy=1)
A = mol.GetNumHeavyAtoms()
alpha = _HallKierAlpha(mol)
denom = P1 + alpha
if denom:
kappa = (A + alpha) * (A + alpha - 1)**2 / denom**2
else:
kappa = 0.0
return round(kappa, 3)
def mmpa_cut(mol: Chem.Mol, rdkit_pattern: bool = False) -> Optional[Set[Any]]:
"""Cut molecules to perform mmpa analysis later
Args:
mol: Molecule to fragment.
rdkit_pattern: Whether to perform the fragmentation
using the default rdkit pattern: [#6+0;!$(*=, #[!#6])]!@!=!#[*]"
Returns:
List of 'smiles,core,chains'
"""
if mol is None:
return mol
outlines = set()
smiles = dm.to_smiles(mol)
if rdkit_pattern:
frags = mmpa_frag(mol, max_cut=3, max_bond_cut=30)
else:
# heavy atoms
frags = mmpa_frag(mol,
pattern="[!#1]!@!=!#[!#1]",
max_cut=4,
max_bond_cut=30)
frags.update(
mmpa_frag(mol,
pattern="[!#1]!@!=!#[!#1]",
max_cut=3,
max_bond_cut=30))
frags = set(frags)
for core, chains in frags:
output = f"{smiles},{core},{chains}\n"
outlines.add(output)
# hydrogen splitting
mol = Chem.AddHs(mol)
smiles = dm.to_smiles(mol)
n = mol.GetNumHeavyAtoms()
if n < 60:
frags = mmpa_frag(mol,
pattern=None,
max_cut=1,
max_bond_cut=100,
h_split=True)
for core, chains in frags:
output = f"{smiles},{core},{chains}\n"
outlines.add(output)
return outlines
def CalculateKappa3(mol: Chem.Mol) -> float:
"""Calculate molecular shape index for three bonded fragments."""
P3 = len(Chem.FindAllPathsOfLengthN(mol, 3))
A = mol.GetNumHeavyAtoms()
denom = P3 + 0.0
if denom:
if A % 2 == 1:
kappa = (A - 1) * (A - 3)**2 / denom**2
else:
kappa = (A - 3) * (A - 2)**2 / denom**2
else:
kappa = 0.0
return round(kappa, 3)
def CalculateHydrophilicityFactor(mol: Chem.Mol) -> float:
"""Calculate hydrophilicity factor.
From Todeschini R. et al., SAR QSAR Environ Res (1997), 7,173-193.
"""
nheavy = mol.GetNumHeavyAtoms()
nc = 0
for atom in mol.GetAtoms():
if atom.GetAtomicNum() == 6:
nc += 1
nhy = 0
for atom in mol.GetAtoms():
if atom.GetAtomicNum() == 7 or atom.GetAtomicNum(
) == 8 or atom.GetAtomicNum() == 16:
atomn = atom.GetNeighbors()
for i in atomn:
if i.GetAtomicNum() == 1:
nhy += 1
res = ((1 + nhy) * math.log((1 + nhy), 2) + nc *
(1.0 / nheavy * math.log(1.0 / nheavy, 2)) + math.sqrt(
(nhy + 0.0) / (nheavy * nheavy))) / math.log(1.0 + nheavy)
return round(res, 3)
def nr_hydrogens(mol: Mol):
return mol.GetNumAtoms() - mol.GetNumHeavyAtoms()
def CalculateHeavyAtomNumber(mol: Chem.Mol) -> float:
"""Calculate number of Heavy atoms."""
return mol.GetNumHeavyAtoms()
