def test_generate_fraggle_fragmentation(self):
mol = Chem.MolFromSmiles('COc1cc(CN2CCC(CC2)NC(=O)c2cncc(C)c2)c(OC)c2ccccc12')
frags = FraggleSim.generate_fraggle_fragmentation(mol)
self.assertEqual(len(frags), 16)
expected = (
'[*]C(=O)NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]c1cc(OC)c2ccccc2c1OC', '[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]Cc1cc(OC)c2ccccc2c1OC',
'[*]Cc1cc(OC)c2ccccc2c1OC.[*]NC(=O)c1cncc(C)c1', '[*]Cc1cc(OC)c2ccccc2c1OC.[*]c1cncc(C)c1',
'[*]NC(=O)c1cncc(C)c1.[*]c1cc(OC)c2ccccc2c1OC', '[*]NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1cncc(C)c1',
'[*]c1c(CN2CCC(NC(=O)c3cncc(C)c3)CC2)cc(OC)c2ccccc12',
'[*]c1c(OC)cc(CN2CCC(NC(=O)c3cncc(C)c3)CC2)c(OC)c1[*]',
'[*]c1cc(CN2CCC(NC(=O)c3cncc(C)c3)CC2)c(OC)c2ccccc12',
'[*]N1CCC(NC(=O)c2cncc(C)c2)CC1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1cncc(C)c1', '[*]c1cc(OC)c2ccccc2c1OC.[*]c1cncc(C)c1')
expected = [_of(s) for s in expected]
for smi in frags:
self.assertTrue(_of(smi) in expected)
# Test case for fragments that contain a cyclic and acyclic component
mol = Chem.MolFromSmiles('c12c(CCC)cccc2cccc1')
frags = FraggleSim.generate_fraggle_fragmentation(mol)
expected = ['[*]CCC.[*]c1ccccc1[*]', '[*]Cc1cccc2ccccc12', '[*]c1cccc(CCC)c1[*]',
'[*]c1cccc2ccccc12', '[*]c1ccccc1[*]']
expected = [_of(s) for s in expected]
for smi in frags:
self.assertTrue(_of(smi) in expected)
def test_GetFraggleSimilarity(self):
q = Chem.MolFromSmiles(
'COc1cc(CN2CCC(NC(=O)c3cncc(C)c3)CC2)c(OC)c2ccccc12')
m = Chem.MolFromSmiles(
'COc1cc(CN2CCC(NC(=O)c3ccccc3)CC2)c(OC)c2ccccc12')
sim, match = FraggleSim.GetFraggleSimilarity(q, m)
self.assertAlmostEqual(sim, 0.980, places=2)
self.assertEqual(match, '[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1')
m = Chem.MolFromSmiles(
'COc1cc(CN2CCC(Nc3nc4ccccc4s3)CC2)c(OC)c2ccccc12')
sim, match = FraggleSim.GetFraggleSimilarity(q, m)
self.assertAlmostEqual(sim, 0.794, places=2)
self.assertEqual(match, '[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1')
q = Chem.MolFromSmiles('COc1ccccc1')
sim, match = FraggleSim.GetFraggleSimilarity(q, m)
self.assertAlmostEqual(sim, 0.347, places=2)
self.assertEqual(match, '[*]c1ccccc1')
m = Chem.MolFromSmiles(
'COc1cc(CN2CCC(NC(=O)c3ccccc3)CC2)c(OC)c2ccccc12')
sim, match = FraggleSim.GetFraggleSimilarity(q, m)
self.assertAlmostEqual(sim, 0.266, places=2)
self.assertEqual(match, '[*]c1ccccc1')
def test_generate_fraggle_fragmentation(self):
mol = Chem.MolFromSmiles('COc1cc(CN2CCC(CC2)NC(=O)c2cncc(C)c2)c(OC)c2ccccc12')
frags = FraggleSim.generate_fraggle_fragmentation(mol)
self.assertEqual(len(frags), 16)
expected = (
'[*]C(=O)NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]c1cc(OC)c2ccccc2c1OC', '[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]Cc1cc(OC)c2ccccc2c1OC',
'[*]Cc1cc(OC)c2ccccc2c1OC.[*]NC(=O)c1cncc(C)c1', '[*]Cc1cc(OC)c2ccccc2c1OC.[*]c1cncc(C)c1',
'[*]NC(=O)c1cncc(C)c1.[*]c1cc(OC)c2ccccc2c1OC', '[*]NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1cncc(C)c1',
'[*]c1c(CN2CCC(NC(=O)c3cncc(C)c3)CC2)cc(OC)c2ccccc12',
'[*]c1c(OC)cc(CN2CCC(NC(=O)c3cncc(C)c3)CC2)c(OC)c1[*]',
'[*]c1cc(CN2CCC(NC(=O)c3cncc(C)c3)CC2)c(OC)c2ccccc12',
'[*]N1CCC(NC(=O)c2cncc(C)c2)CC1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1cncc(C)c1', '[*]c1cc(OC)c2ccccc2c1OC.[*]c1cncc(C)c1')
expected = [_of(s) for s in expected]
for smi in frags:
self.assertTrue(_of(smi) in expected)
# Test case for fragments that contain a cyclic and acyclic component
mol = Chem.MolFromSmiles('c12c(CCC)cccc2cccc1')
frags = FraggleSim.generate_fraggle_fragmentation(mol)
expected = ['[*]CCC.[*]c1ccccc1[*]', '[*]Cc1cccc2ccccc12', '[*]c1cccc(CCC)c1[*]',
'[*]c1cccc2ccccc12', '[*]c1ccccc1[*]']
expected = [_of(s) for s in expected]
for smi in frags:
self.assertTrue(_of(smi) in expected)
def test_isValidRingCut(self):
rdBase.DisableLog('rdApp.error')
self.assertEqual(FraggleSim.isValidRingCut(Chem.MolFromSmiles('[*]CCC[*]')), False)
self.assertEqual(FraggleSim.isValidRingCut(Chem.MolFromSmiles('[*]C1CC1[*]')), True)
self.assertEqual(FraggleSim.isValidRingCut(Chem.MolFromSmiles('[*]c1ccccc1[*]')), True)
self.assertEqual(
FraggleSim.isValidRingCut(Chem.MolFromSmiles('[*]cccc[*]', sanitize=False)), False)
rdBase.EnableLog('rdApp.error')
def test_isValidRingCut(self):
rdBase.DisableLog('rdApp.error')
self.assertEqual(FraggleSim.isValidRingCut(Chem.MolFromSmiles('[*]CCC[*]')), False)
self.assertEqual(FraggleSim.isValidRingCut(Chem.MolFromSmiles('[*]C1CC1[*]')), True)
self.assertEqual(FraggleSim.isValidRingCut(Chem.MolFromSmiles('[*]c1ccccc1[*]')), True)
self.assertEqual(
FraggleSim.isValidRingCut(Chem.MolFromSmiles('[*]cccc[*]', sanitize=False)), False)
rdBase.EnableLog('rdApp.error')
def testFragmentation(self):
"""
"""
mol = Chem.MolFromSmiles(
'COc1cc(CN2CCC(CC2)NC(=O)c2cncc(C)c2)c(OC)c2ccccc12')
frags = FraggleSim.generate_fraggle_fragmentation(mol)
self.assertEqual(len(frags), 16)
expected = ('[*]C(=O)NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]Cc1cc(OC)c2ccccc2c1OC',
'[*]Cc1cc(OC)c2ccccc2c1OC.[*]NC(=O)c1cncc(C)c1',
'[*]Cc1cc(OC)c2ccccc2c1OC.[*]c1cncc(C)c1',
'[*]NC(=O)c1cncc(C)c1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1cncc(C)c1',
'[*]c1c(CN2CCC(NC(=O)c3cncc(C)c3)CC2)cc(OC)c2ccccc12',
'[*]c1c(OC)cc(CN2CCC(NC(=O)c3cncc(C)c3)CC2)c(OC)c1[*]',
'[*]c1cc(CN2CCC(NC(=O)c3cncc(C)c3)CC2)c(OC)c2ccccc12',
'[*]N1CCC(NC(=O)c2cncc(C)c2)CC1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1cncc(C)c1',
'[*]c1cc(OC)c2ccccc2c1OC.[*]c1cncc(C)c1')
for smi in frags:
self.assertTrue(smi in expected)
def testFragmentation(self):
"""
"""
mol = Chem.MolFromSmiles('COc1cc(CN2CCC(CC2)NC(=O)c2cncc(C)c2)c(OC)c2ccccc12')
frags = FraggleSim.generate_fraggle_fragmentation(mol)
self.failUnlessEqual(len(frags),16)
expected=('[*]C(=O)NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1cncc(C)c1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]Cc1cc(OC)c2ccccc2c1OC.[*]C(=O)c1cncc(C)c1',
'[*]Cc1cc(OC)c2ccccc2c1OC.[*]NC(=O)c1cncc(C)c1',
'[*]Cc1cc(OC)c2ccccc2c1OC.[*]c1cncc(C)c1',
'[*]NC(=O)c1cncc(C)c1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1cncc(C)c1',
'[*]c1c(CN2CCC(NC(=O)c3cncc(C)c3)CC2)cc(OC)c2ccccc12',
'[*]c1c([*])c(OC)c(CN2CCC(NC(=O)c3cncc(C)c3)CC2)cc1OC',
'[*]c1cc(CN2CCC(NC(=O)c3cncc(C)c3)CC2)c(OC)c2ccccc12',
'[*]c1cc(OC)c2ccccc2c1OC.[*]N1CCC(NC(=O)c2cncc(C)c2)CC1',
'[*]c1cncc(C)c1.[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]c1cncc(C)c1.[*]c1cc(OC)c2ccccc2c1OC')
for smi in frags:
self.failUnless(smi in expected)
def frag(
mol: Chem.Mol,
remove_parent: bool = False,
sanitize: bool = True,
fix: bool = True,
):
"""Generate all possible fragmentation of a molecule.
Args:
mol: a molecule.
remove_parent: Remove parent from the fragments.
sanitize: Wether to sanitize the fragments.
fix: Wether to fix the fragments.
"""
frags = FraggleSim.generate_fraggle_fragmentation(mol)
smiles = set([])
for seq in frags:
smiles |= {s.strip() for s in seq.split(".")}
smiles = list(sorted(smiles, reverse=True))
frags = [dm.to_mol(s) for s in smiles]
if fix:
frags = [dm.fix_mol(x) for x in frags]
if sanitize:
frags = [dm.sanitize_mol(x) for x in frags]
frags = [x for x in frags if x is not None]
if remove_parent:
return frags
return [mol] + frags
def generate_sim(base: str, smile: str):
base = Chem.MolFromSmiles(base)
smile = Chem.MolFromSmiles(smile)
try:
sim, match = FraggleSim.GetFraggleSimilarity(smile, base)
except:
return 0.0, None
return sim, match
def testFragmentation2(self):
mol = Chem.MolFromSmiles('COc1cc(CN2CCC(NC(=O)c3ccccc3)CC2)c(OC)c2ccccc12')
frags = FraggleSim.generate_fraggle_fragmentation(mol)
self.assertEqual(len(frags), 13)
expected = (
'[*]C(=O)c1ccccc1.[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1ccccc1.[*]Cc1cc(OC)c2ccccc2c1OC', '[*]C(=O)c1ccccc1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1ccccc1',
'[*]Cc1cc(OC)c2ccccc2c1OC.[*]NC(=O)c1ccccc1', '[*]Cc1cc(OC)c2ccccc2c1OC.[*]c1ccccc1',
'[*]N1CCC(NC(=O)c2ccccc2)CC1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]NC(=O)c1ccccc1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1ccccc1',
'[*]c1c(CN2CCC(NC(=O)c3ccccc3)CC2)cc(OC)c2ccccc12',
'[*]c1c(OC)cc(CN2CCC(NC(=O)c3ccccc3)CC2)c(OC)c1[*]',
'[*]c1cc(CN2CCC(NC(=O)c3ccccc3)CC2)c(OC)c2ccccc12', '[*]c1cc(OC)c2ccccc2c1OC.[*]c1ccccc1')
expected = [_of(s) for s in expected]
for smi in frags:
self.assertIn(_of(smi), expected)
def testFragmentation2(self):
mol = Chem.MolFromSmiles('COc1cc(CN2CCC(NC(=O)c3ccccc3)CC2)c(OC)c2ccccc12')
frags = FraggleSim.generate_fraggle_fragmentation(mol)
self.assertEqual(len(frags), 13)
expected = (
'[*]C(=O)c1ccccc1.[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1',
'[*]C(=O)c1ccccc1.[*]Cc1cc(OC)c2ccccc2c1OC', '[*]C(=O)c1ccccc1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]C1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1ccccc1',
'[*]Cc1cc(OC)c2ccccc2c1OC.[*]NC(=O)c1ccccc1', '[*]Cc1cc(OC)c2ccccc2c1OC.[*]c1ccccc1',
'[*]N1CCC(NC(=O)c2ccccc2)CC1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]NC(=O)c1ccccc1.[*]c1cc(OC)c2ccccc2c1OC',
'[*]NC1CCN(Cc2cc(OC)c3ccccc3c2OC)CC1.[*]c1ccccc1',
'[*]c1c(CN2CCC(NC(=O)c3ccccc3)CC2)cc(OC)c2ccccc12',
'[*]c1c(OC)cc(CN2CCC(NC(=O)c3ccccc3)CC2)c(OC)c1[*]',
'[*]c1cc(CN2CCC(NC(=O)c3ccccc3)CC2)c(OC)c2ccccc12', '[*]c1cc(OC)c2ccccc2c1OC.[*]c1ccccc1')
expected = [_of(s) for s in expected]
for smi in frags:
self.assertIn(_of(smi), expected)
if __name__ == '__main__':
import sys, re
if (len(sys.argv) >= 2):
print(
"Program to run the first part of Fraggle. Program splits the molecule\nready for the search\n")
print("USAGE: ./fraggle.py <file_of_smiles")
print("Format of smiles file: SMILES ID (space or comma separated)")
print("Output: whole mol smiles,ID,fraggle split smiles\n")
sys.exit(1)
#read the STDIN
for line in sys.stdin:
line = line.rstrip()
smi, id_ = re.split('\s|,', line)
#print smi,id_
mol = Chem.MolFromSmiles(smi)
if mol is None:
sys.stderr.write("Can't generate mol for: %s\n" % (smi))
continue
out_fragments = FraggleSim.generate_fraggle_fragmentation(mol)
#print out the unique fragments
for x in out_fragments:
#cansmi
temp = Chem.MolFromSmiles(x)
print("%s,%s,%s" % (smi, id_, Chem.MolToSmiles(temp)))
# test_mol_neu.SetProp("NScode", mol.GetProp("NScode"))
reference_desalt.append(test_mol_neu)
count = count+1
print("Reference Molecule Num: ", len(reference_desalt))
# Morganによる類似性評価
morgan_reference = [AllChem.GetMorganFingerprintAsBitVect(mol, 2, 2048) for mol in reference_desalt] #(mol,中心からの半径,bit数)
morgan_query = [AllChem.GetMorganFingerprintAsBitVect(mol, 2, 2048) for mol in query]
tanimoto = DataStructs.BulkTanimotoSimilarity(morgan_query[0], morgan_reference)
# Fraggleによる類似性評価
fraggle_sim = []
fraggle_match = []
for (sim, match) in [FraggleSim.GetFraggleSimilarity(query[0], reference_desalt[i]) for i in range(len(reference_desalt))] :
fraggle_sim.append(sim)
fraggle_match.append(match)
# 結果をPDFファイルに書き込み
# img = Draw.MolsToGridImage(namiki[:16], molsPerRow=4, subImgSize=(300,200), legends=['Fraggle: {:.2f}'.format(i) for i in fraggle_similarity[:16]])
# img.save('./fraggle.pdf')
# 結果をSDFファイルに書き込み
sdf_filename = "./fraggle-result.sdf"
writer = Chem.SDWriter(sdf_filename)
query[0].SetProp("NScode", "2-17") #molオブジェクトにプロパティを付加する
query[0].SetProp("Tanimoto_Sim", "1")
query[0].SetProp("Fraggle_Sim", "1")
writer.write(query[0])
if (iMol == None):
sys.stderr.write("Can't generate mol for: %s\n" % (inSmi))
continue
#discard based on atom size
if (iMol.GetNumAtoms() < query_size[qID] - 3):
#sys.stderr.write("Too small: %s\n" % (inSmi) )
continue
if (iMol.GetNumAtoms() > query_size[qID] + 4):
#sys.stderr.write("Too large: %s\n" % (inSmi) )
continue
#print '>>>',id_
rdkit_sim, fraggle_sim = FraggleSim.compute_fraggle_similarity_for_subs(
iMol, query_mols[qID], qSmi, qSubs, options.pfp)
day_sim[qID][id_] = rdkit_sim
frag_sim[qID][id_] = max(frag_sim[qID][id_], fraggle_sim)
#check if you have the fp for the modified query
#and generate if need to
#right, print out the results for the query
#Format: SMILES,ID,QuerySMI,QueryID,Fraggle_Similarity,Daylight_Similarity
for qID in frag_sim:
for id_ in frag_sim[qID]:
if (frag_sim[qID][id_] >= fraggle_cutoff):
print("%s,%s,%s,%s,%s,%s" %
(id_to_smi[id_], id_, id_to_smi[qID], qID, frag_sim[qID][id_], day_sim[qID][id_]))
if(iMol == None):
sys.stderr.write("Can't generate mol for: %s\n" % (inSmi) )
continue
#discard based on atom size
if(iMol.GetNumAtoms() < query_size[qID]-3):
#sys.stderr.write("Too small: %s\n" % (inSmi) )
continue;
if(iMol.GetNumAtoms() > query_size[qID]+4):
#sys.stderr.write("Too large: %s\n" % (inSmi) )
continue;
#print '>>>',id_
rdkit_sim,fraggle_sim = FraggleSim.compute_fraggle_similarity_for_subs(iMol,query_mols[qID],qSmi,qSubs,
options.pfp)
day_sim[qID][id_] = rdkit_sim
frag_sim[qID][id_] = max(frag_sim[qID][id_],fraggle_sim)
#check if you have the fp for the modified query
#and generate if need to
#right, print out the results for the query
#Format: SMILES,ID,QuerySMI,QueryID,Fraggle_Similarity,Daylight_Similarity
for qID in frag_sim:
for id_ in frag_sim[qID]:
if(frag_sim[qID][id_] >= fraggle_cutoff):
print("%s,%s,%s,%s,%s,%s" % (id_to_smi[id_],id_,id_to_smi[qID],qID,frag_sim[qID][id_],day_sim[qID][id_]))
if __name__ =='__main__':
import sys,re
if (len(sys.argv) >= 2):
print "Program to run the first part of Fraggle. Program splits the molecule\nready for the search\n"
print "USAGE: ./fraggle.py <file_of_smiles"
print "Format of smiles file: SMILES ID (space or comma separated)"
print "Output: whole mol smiles,ID,fraggle split smiles\n"
sys.exit(1)
#read the STDIN
for line in sys.stdin:
line = line.rstrip()
smi,id_ = re.split('\s|,',line)
#print smi,id_
mol = Chem.MolFromSmiles(smi)
if mol is None:
sys.stderr.write("Can't generate mol for: %s\n" % (smi) )
continue
out_fragments = FraggleSim.generate_fraggle_fragmentation(mol)
#print out the unique fragments
for x in out_fragments:
#cansmi
temp = Chem.MolFromSmiles(x)
print "%s,%s,%s" % (smi,id_,Chem.MolToSmiles(temp))
