def test_ecfp(self):
"""Test Butyramide (example used in ECFP paper)"""
mol = smilin("CCCC(=O)N")
# ECFP_0
fp = ecfp(mol, radius=0)
assert len(fp) == 5
# ECFP_2
fp = ecfp(mol, radius=1)
assert len(fp) == 11
# ECFP_4
fp = ecfp(mol, radius=2)
assert len(fp) == 14
# ECFP_6
fp = ecfp(mol, radius=3)
assert len(fp) == 14
def fingerprint(self):
"""
Compute the fingerprint for a compound
:returns: The fingerprint sparse vector dict
"""
try:
return getattr(self, '_fp')
except AttributeError:
pass
self._fp = {}
if self.mol() is not None:
self._fp = ecfp(self.mol())
return self._fp
def molecule_embedding(filename):
dict = {}
df = pd.read_excel(filename)
for id, smile in zip(df.MOLENAME, df.SMILES):
try:
mol = smilin(smile)
try:
fp = ecfp(mol, radius=3)
n_bits = 1024 # Number of bits in fixed-length fingerprint
fingerprint = [0 for _ in range(n_bits)
] #fingerprints as a python list
for nbrhood_hash in fp.keys():
bit = nbrhood_hash % n_bits
fingerprint[bit] = 1
#print(fingerprint)
dict[id] = fingerprint
except:
print('Bad ecfp for', id)
except:
print('Bad mol for', id)
return list(dict.values())
