def PrintTanimoto(molA, molB, minb, maxb):
fpA = oegraphsim.OEFingerPrint()
fpB = oegraphsim.OEFingerPrint()
numbits = 2048
atype = oegraphsim.OEFPAtomType_DefaultPathAtom
btype = oegraphsim.OEFPBondType_DefaultPathBond
oegraphsim.OEMakePathFP(fpA, molA, numbits, minb, maxb, atype, btype)
oegraphsim.OEMakePathFP(fpB, molB, numbits, minb, maxb, atype, btype)
print("Tanimoto(A,B) = %.3f" % oegraphsim.OETanimoto(fpA, fpB))
def PrintTanimoto(molA, molB, atype, btype):
fpA = oegraphsim.OEFingerPrint()
fpB = oegraphsim.OEFingerPrint()
numbits = 2048
minb = 0
maxb = 5
oegraphsim.OEMakePathFP(fpA, molA, numbits, minb, maxb, atype, btype)
oegraphsim.OEMakePathFP(fpB, molB, numbits, minb, maxb, atype, btype)
print("Tanimoto(A,B) = %.3f" % oegraphsim.OETanimoto(fpA, fpB))
def calculate_mol_similarity(molA, molB):
"""
Function to calculate the similarity between two oemol objects
should be used to utils/openeye.py or openmoltools
:param molA: oemol object of molecule A
:param molB: oemol object of molecule B
:return: float, tanimoto score of the two molecules, between 0 and 1
"""
fpA = oegraphsim.OEFingerPrint()
fpB = oegraphsim.OEFingerPrint()
oegraphsim.OEMakeFP(fpA, molA, oegraphsim.OEFPType_MACCS166)
oegraphsim.OEMakeFP(fpB, molB, oegraphsim.OEFPType_MACCS166)
return oegraphsim.OETanimoto(fpA, fpB)
def calculate_fp(self):
for mol in self.act_list:
fp = oegraphsim.OEFingerPrint()
oegraphsim.OEMakeFP(fp, mol, self.args.fptype)
self.fp_list.append(fp)
def circular_wrapper(smi, num_bits=1024, min_radius=2, max_radius=2):
mol = oechem.OEGraphMol()
oechem.OESmilesToMol(mol, smi)
fp = oegraphsim.OEFingerPrint()
oegraphsim.OEMakeCircularFP(fp, mol, num_bits, min_radius, max_radius,
oegraphsim.OEFPAtomType_DefaultPathAtom,
oegraphsim.OEFPBondType_DefaultPathBond)
return get_bit_string(fp)
def maccs_wrapper(smi, ):
mol = oechem.OEGraphMol()
oechem.OESmilesToMol(mol, smi)
fp = oegraphsim.OEFingerPrint()
oegraphsim.OEMakeMACCS166FP(
fp,
mol,
)
return get_bit_string(fp)
# Customer is hereby permitted to use, copy, and modify the Sample Code,
# subject to these terms. OpenEye claims no rights to Customer's
# modifications. Modification of Sample Code is at Customer's sole and
# exclusive risk. Sample Code may require Customer to have a then
# current license or subscription to the applicable OpenEye offering.
# THE SAMPLE CODE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED. OPENEYE DISCLAIMS ALL WARRANTIES, INCLUDING, BUT
# NOT LIMITED TO, WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
# PARTICULAR PURPOSE AND NONINFRINGEMENT. In no event shall OpenEye be
# liable for any damages or liability in connection with the Sample Code
# or its use.
from openeye import oechem
from openeye import oegraphsim
mol = oechem.OEGraphMol()
oechem.OESmilesToMol(mol, "c1ccccc1")
# @ <SNIPPET-MAKE-FP>
fpA = oegraphsim.OEFingerPrint()
oegraphsim.OEMakePathFP(fpA, mol)
fpB = oegraphsim.OEFingerPrint()
oegraphsim.OEMakeFP(fpB, mol, fpA.GetFPTypeBase())
# @ </SNIPPET-MAKE-FP>
if oegraphsim.OEIsSameFPType(fpA, fpB):
print("same fingerprint types")
else:
print("different fingerprint types")
# current license or subscription to the applicable OpenEye offering. # THE SAMPLE CODE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, # EXPRESS OR IMPLIED. OPENEYE DISCLAIMS ALL WARRANTIES, INCLUDING, BUT # NOT LIMITED TO, WARRANTIES OF MERCHANTABILITY, FITNESS FOR A # PARTICULAR PURPOSE AND NONINFRINGEMENT. In no event shall OpenEye be # liable for any damages or liability in connection with the Sample Code # or its use. from openeye import oechem from openeye import oegraphsim mol = oechem.OEGraphMol() oechem.OESmilesToMol(mol, "c1ccccc1") # @ <SNIPPET-MAKE-PATH-1> fp = oegraphsim.OEFingerPrint() oegraphsim.OEMakePathFP(fp, mol) # @ </SNIPPET-MAKE-PATH-1> print(fp.GetFPTypeBase().GetFPTypeString()) # @ <SNIPPET-MAKE-PATH-2> oegraphsim.OEMakeFP(fp, mol, oegraphsim.OEFPType_Path) # @ </SNIPPET-MAKE-PATH-2> print(fp.GetFPTypeBase().GetFPTypeString()) # @ <SNIPPET-MAKE-PATH-3> numbits = 1024 minbonds = 0 maxbonds = 5
def cluster_similar_molecules(
smiles, fingerprint_type=oegraphsim.OEFPType_Tree, eps=0.5, min_samples=2
):
"""The method attempts to cluster a sets of molecules based on their
similarity using a Tanimoto distance metric and the `sklearn` DBSCAN
clustering code.
Notes
-----
This is based on the code by David Mobley:
https://github.com/openforcefield/release-1-benchmarking/blob/master/QM_molecule_selection/divide_sets.ipynb
Parameters
----------
smiles: list of str
The SMILES representations of the molecules to cluster.
fingerprint_type
The type of molecular fingerprint to use.
eps: float
The `eps` parameter to pass as an argument to DBSCAN while clustering.
min_samples: int
The `min_samples` parameter to pass as an argument to DBSCAN while
clustering.
Returns
-------
dict of str and list of str
The clustered SMILES patterns.
"""
assert isinstance(smiles, list)
# Build fingerprints
fingerprints = {}
for smiles_pattern in smiles:
oe_molecule = oechem.OEMol()
oechem.OEParseSmiles(oe_molecule, smiles_pattern)
fingerprint = oegraphsim.OEFingerPrint()
oegraphsim.OEMakeFP(fingerprint, oe_molecule, fingerprint_type)
fingerprints[smiles_pattern] = fingerprint
# Build a similarity matrix
distance_matrix = np.zeros((len(smiles), len(smiles)))
for i, smiles_i in enumerate(smiles):
for j, smiles_j in enumerate(smiles):
if i == j:
continue
distance_matrix[i, j] = 1.0 - oegraphsim.OETanimoto(
fingerprints[smiles_i], fingerprints[smiles_j]
)
# Cluster the data
clustering = DBSCAN(eps=eps, min_samples=min_samples, metric="precomputed")
clustered_smiles = clustering.fit(distance_matrix)
labels = clustered_smiles.labels_
smiles_by_cluster = {}
for label in set(labels):
smiles_by_cluster[label] = [
smiles[i] for i, x in enumerate(labels) if x == label
]
return smiles_by_cluster
def main(argv=[__name__]):
itf = oechem.OEInterface()
oechem.OEConfigure(itf, InterfaceData)
defopts = oegraphsim.OEFPDatabaseOptions(10,
oegraphsim.OESimMeasure_Tanimoto)
oegraphsim.OEConfigureFPDatabaseOptions(itf, defopts)
oegraphsim.OEConfigureFingerPrint(
itf, oegraphsim.OEGetFPType(oegraphsim.OEFPType_Tree))
if not oechem.OEParseCommandLine(itf, argv):
return 0
qfname = itf.GetString("-query")
mfname = itf.GetString("-molfname")
ofname = itf.GetString("-out")
# initialize databases
timer = oechem.OEWallTimer()
timer.Start()
ifs = oechem.oemolistream()
if not ifs.open(qfname):
oechem.OEThrow.Fatal("Cannot open input file!")
query = oechem.OEGraphMol()
if not oechem.OEReadMolecule(ifs, query):
oechem.OEThrow.Fatal("Cannot read query molecule!")
moldb = oechem.OEMolDatabase()
if not moldb.Open(mfname):
oechem.OEThrow.Fatal("Cannot open molecule database!")
ofs = oechem.oemolostream()
if not ofs.open(ofname):
oechem.OEThrow.Fatal("Cannot open output file!")
fptype = oegraphsim.OESetupFingerPrint(itf)
oechem.OEThrow.Info("Using fingerprint type %s" % fptype.GetFPTypeString())
fpdb = oegraphsim.OEFPDatabase(fptype)
emptyfp = oegraphsim.OEFingerPrint()
emptyfp.SetFPTypeBase(fptype)
nrmols = moldb.GetMaxMolIdx()
mol = oechem.OEGraphMol()
for idx in range(0, nrmols):
if moldb.GetMolecule(mol, idx):
fpdb.AddFP(mol)
else:
fpdb.AddFP(emptyfp)
nrfps = fpdb.NumFingerPrints()
oechem.OEThrow.Info("%5.2f sec to initialize databases" % timer.Elapsed())
opts = oegraphsim.OEFPDatabaseOptions()
oegraphsim.OESetupFPDatabaseOptions(opts, itf)
# search fingerprint database
timer.Start()
scores = fpdb.GetSortedScores(query, opts)
oechem.OEThrow.Info("%5.2f sec to search %d fingerprints" %
(timer.Elapsed(), nrfps))
timer.Start()
hit = oechem.OEGraphMol()
for si in scores:
if moldb.GetMolecule(hit, si.GetIdx()):
oechem.OEWriteMolecule(ofs, hit)
oechem.OEThrow.Info("%5.2f sec to write %d hits" %
(timer.Elapsed(), opts.GetLimit()))
return 0
