def match(smiles):
group_ids = []
if smiles.find('.') > -1: # ignore complex
return group_ids
py_mol = pybel.readstring('smi', smiles)
py_mol.removeh()
formula = py_mol.formula
atom_set = set(Formula.read(formula).atomdict.keys())
if not atom_set <= {'C', 'H', 'O', 'N', 'F', 'Cl', 'Br'}:
return group_ids
if not ('C' in atom_set and {'H', 'F', 'Cl', 'Br'} & atom_set != set()):
return group_ids
if atom_set == {'C', 'H'}:
group_ids.append(smarts_id['hydrocarbon'])
for s in ['[CX2]', '[CX3]', 'c', '[#6;v0,v1,v2,v3]']:
if pybel.Smarts(s).findall(py_mol) != []:
break
else:
group_ids.append(smarts_id['alkane'])
if pybel.Smarts('[OH]').findall(py_mol).__len__() > 1:
group_ids.append(smarts_id['diol'])
for name, smarts in smarts_dict.items():
if pybel.Smarts(smarts).findall(py_mol):
group_ids.append(smarts_id[smarts])
return group_ids
def index(self, smiles):
# bridged atoms
bridg_Matcher = pybel.Smarts('[x3]')
# spiro atoms
spiro_Matcher = pybel.Smarts('[x4]')
# linked rings
RR_Matcher = pybel.Smarts('[R]!@[R]')
# separated rings
R_R_Matcher = pybel.Smarts('[R]!@*!@[R]')
rd_mol: Mol = Chem.MolFromSmiles(smiles)
py_mol = pybel.readstring('smi', smiles)
index = [
py_mol.OBMol.NumHvyAtoms(),
int(round(py_mol.molwt, 1) * 10),
self.get_shortest_wiener(rd_mol)[0],
Chem.CalcNumRotatableBonds(Chem.AddHs(rd_mol)),
len(bridg_Matcher.findall(py_mol)),
len(spiro_Matcher.findall(py_mol)),
len(RR_Matcher.findall(py_mol)),
len(R_R_Matcher.findall(py_mol)),
] + \
list(self.get_ring_info(py_mol))
return np.array(index)
def run():
inputfile=pybel.readfile(sys.argv[1].split(".")[-1],sys.argv[1])
value=()
for mol in inputfile:
descvalues=mol.calcdesc()
value= value+(descvalues.get('TPSA'),)
value= value+(descvalues.get('HBD'),)
value= value+(descvalues.get('logP'),)
value= value+(descvalues.get('MW'),)
value= value+(descvalues.get('tbonds'),)
value= value+(descvalues.get('nF'),)
value= value+(descvalues.get('bonds'),)
value= value+(descvalues.get('atoms'),)
value= value+(descvalues.get('HBA1'),)
value= value+(descvalues.get('HBA2'),)
value= value+(descvalues.get('sbonds'),)
value= value+(descvalues.get('dbonds'),)
value= value+(descvalues.get('MR'),)
value= value+(descvalues.get('abonds'),)
smarts = pybel.Smarts("[+]")
num=smarts.findall(mol)
value= value+(len(num),)
smarts = pybel.Smarts("[-]")
num=smarts.findall(mol)
value= value+(len(num),)
model=joblib.load('volume_model/volume.pkl')
for result in model.predict(value):
return round(result,2)
def get_properties_ext(mol):
HBD = pybel.Smarts("[!#6;!H0]")
HBA = pybel.Smarts("[$([$([#8,#16]);!$(*=N~O);" +
"!$(*~N=O);X1,X2]),$([#7;v3;" +
"!$([nH]);!$(*(-a)-a)])]")
calc_desc_dict = mol.calcdesc()
try:
logp = calc_desc_dict['logP']
except:
logp = calc_desc_dict['LogP']
return {
"molwt": mol.molwt,
"logp": logp,
"donors": len(HBD.findall(mol)),
"acceptors": len(HBA.findall(mol)),
"psa": calc_desc_dict['TPSA'],
"mr": calc_desc_dict['MR'],
"rotbonds": mol.OBMol.NumRotors(),
"can": mol.write("can").split()[0].strip(
), ### tthis one works fine for both zinc and chembl (no ZINC code added after can descriptor string)
"inchi": mol.write("inchi").strip(),
"inchi_key": get_inchikey(mol).strip(),
"rings": len(mol.sssr),
"atoms": mol.OBMol.NumHvyAtoms(),
"spectrophore": OBspectrophore(mol),
}
def main(substrate, group, sub_id, group_id, position, ref):
# ========================> Generate Smiles < ========================= #
if not group:
smiles = substrate
else:
if position == 1:
# ============= Put group at the ortho positions ============== #
smiles = substrate[0:2] + "(" + group + ")" + substrate[2:]
elif position == 2:
# ============= Put group at the meta positions =============== #
smiles = substrate[0:3] + "(" + group + ")" + substrate[3:]
elif position == 3:
# ============= Put group at the para positions =============== #
smiles = substrate[0:4] + "(" + group + ")" + substrate[4:]
print(smiles)
# ============== Generate Folder Name ========== #
folder = ip.iupac_name(smiles, substrate, group, sub_id, group_id,
position)
print folder
os.system('mkdir %s' % folder)
os.chdir(folder)
os.system('mkdir scratch')
# ============== Copy data to folder =========== #
#TODO
# ============== Generate Molecule ========== #
mol = confab.gen3d(smiles)
# ============ Driving Coordinate Idx Generation ==============#
smarts1 = pybel.Smarts(substrate)
smarts2 = pybel.Smarts("cccccc")
r1 = smarts1.findall(mol)
r2 = smarts2.findall(mol)
#============ confab and align to reference ==================#
r1 = [r1[0][-2], r1[0][-1]]
r2 = list(r2[0])
mol = confab.confab(mol, r1, r2)
# => align to substructure <=#
path = os.getcwd() + "/tmp.xyz"
if ref is not None:
cmd_str = "obabel %s %s -O %s -s %s --align" % (ref, path, path,
substrate)
print cmd_str
os.system(cmd_str)
#read aligned geom
mol = confab.read_molecules(path, single=False)
mol = mol[1]
#============ do zstruct ===================================== #
zstruct2.zstruct1(mol, r1, r2, None, doOne=False, doTwo=True)
def __init__(self, ec, substrate, product):
self.ec = ec
self.substrate = substrate
self.product = product
try:
self.smarts_subs = pybel.Smarts(substrate)
self.smarts_prod = pybel.Smarts(product)
except IOError:
logging.warning("failed parsing: %s >> %s" % (substrate, product))
self.smarts_subs = None
self.smarts_prod = None
def teste():
phenol = pybel.Smarts("[OH]c1ccccc1") # phenol
ethyl = pybel.Smarts("[#6][#6]") # ethyl group
# mol = pybel.readfile("mol2", "resources/base.mol2")
found_total = 0
found = []
for mol in pybel.readfile("mol2", "resources/base.mol2"):
finded = phenol.findall(mol)
found_total += len(finded)
found.append(finded)
return mol.draw(True)
def count(self, smilesstr):
##
groups, include, evalkw, brackets, quotes, unquote = self.commonattr()
haskw, hasbracket, hasquote = self.matchedpatt(groups)
if include is None:
include = [True] * len(groups)
##
mol = pybel.readstring('smi', smilesstr)
mol.addh()
molecule = mol # copy reference; keyword for userdef.py 29.09.2015
abundances = pd.Series([np.nan] * len(groups), index=groups.index)
## SMARTS search
for key in groups.index[~haskw & ~hasbracket & ~hasquote]:
abundances[key] = len(pybel.Smarts(groups[key]).findall(mol))
## evaluate eval keyword
for key in groups.index[haskw]: # untested
abundances[key] = round(eval(evalkw.search(groups[key]).group(1)))
## evaluated quoted expressions
for key in groups.index[hasquote]: # untested
abundances[key] = round(eval(
unquote(groups[key]).format(**groups)))
## evaluate expressions
orderedexpr = self.__orderexpr(groups, hasbracket, brackets)
for key in orderedexpr: #groups.index[hasbracket]:
abundances[key] = round(eval(groups[key].format(**abundances)))
##
return abundances[include].astype(int)
def pattern():
global SMARTS
PATTERNS = []
for smarts in SMARTS:
el = pybel.Smarts(smarts)
PATTERNS.append(el)
return PATTERNS
def integer_fp(smile, smartsqfile='SMARTSFileFull.json'):
'''Creates an integer-valued fingerprint. Probably takes longer since it is a VERY hacked version, but it generates
data for SVMs.'''
rdmol = Chem.MolFromSmiles(smile)
if rdmol:
smartsData = open(os.path.join(CHEMPATH, smartsqfile))
encodedSmartsDict = json.load(smartsData)
smartsData.close()
smartsDict = {k.encode('utf-8'): v.encode('utf-8').split(': ')[1].strip(' \n') for k, v in encodedSmartsDict.iteritems()}
fpt = [0] * len(smartsDict.keys())
cpd = pybel.readstring('smi', smile)
for k, v in smartsDict.iteritems():
query = pybel.Smarts(v)
fpt[int(k) - 1] = len(query.findall(cpd))
#SMARTS querying directly for chirality does not give RS isomerism, so adding it at the end. May get flexed to
#_check_reverse for consistency.
rect, sini = chem.flag_chiral(rdmol)
fpt.append(rect)
fpt.append(sini)
return fpt
else:
return None
def get_abundances(self, smilesstr=None):
## main body
mol = pybel.readstring('smi', smilesstr)
## store SIMPOL patterns in ordered dictionary
abundances = OrderedDict()
for key, patt in self.smartspatt.items()[0:]:
abundances[key] = 0 if 'nomatch' in patt else \
len(pybel.Smarts(patt).findall(mol))
## find auxiliary patterns
aux = OrderedDict()
for key, patt in self.smartsaux.items():
aux[key] = len(pybel.Smarts(patt).findall(mol))
## combine and return pandas series
abundances[
'H'] = aux['h1'] + 2 * aux['h2'] + 3 * aux['h3'] + 4 * aux['h4']
return pd.Series(abundances)
def matchatoms(self, smilesstr):
##
groups, include, evalkw, brackets, quotes, unquote = self.commonattr()
haskw, hasbracket, hasquote = self.matchedpatt(groups)
##
mol = pybel.readstring('smi', smilesstr)
mol.addh()
molecule = mol # copy reference; keyword for userdef.py 29.09.2015
tups = OrderedDict(zip(groups.index, [None] * len(groups)))
## SMARTS search
for key in groups.index[~haskw & ~hasbracket & ~hasquote]:
tups[key] = set(pybel.Smarts(groups[key]).findall(mol))
## evaluate eval keyword
for key in groups.index[haskw]:
tups[key] = eval(evalkw.search(groups[key]).group(1))
## evaluate quoted expressions
for key in groups.index[hasquote]: # untested
tups[key] = self.__substitute(mol, groups[key], quotes, groups)
## evaluate expressions
orderedexpr = self.__orderexpr(groups, hasbracket, brackets)
for key in orderedexpr: #groups.index[hasbracket]
tups[key] = self.__substitute(mol, groups[key], brackets, tups)
usetups = OrderedDict([(k, v) for (k, v) in tups.items()
if include.ix[k]])
alltups = reduce(set.union, usetups.values())
allatoms = reduce(add, map(list, alltups), [])
atomicmass = set([(atom.type, atom.atomicmass) for atom in mol.atoms
if atom.idx in allatoms])
##
idxlabel = 'atom'
atomtype = pd.DataFrame([(atom.idx, atom.type) for atom in mol.atoms],
columns=[idxlabel, 'type']).set_index(idxlabel)
matched = self.__atomtable(atomtype, usetups)
##
return (matched, atomicmass)
def detect_dihedrals(mol: pybel.Molecule) -> List[DihedralInfo]:
"""Detect the bonds to be treated as rotors.
We use the more generous definition from RDKit:
https://github.com/rdkit/rdkit/blob/1bf6ef3d65f5c7b06b56862b3fb9116a3839b229/rdkit/Chem/Lipinski.py#L47%3E
It matches pairs of atoms that are connected by a single bond,
both bonds have at least one other bond that is not a triple bond
and they are not part of the same ring.
Args:
mol: Molecule to assess
Returns:
List of dihedral angles. Most are defined
"""
dihedrals = []
# Compute the bonding graph
g = get_bonding_graph(mol)
# Get the indices of backbond atoms
backbone = set(i for i, d in g.nodes(data=True) if d['z'] > 1)
# Step 1: Get the bonds from a simple matching
smarts = pybel.Smarts('[!$(*#*)&!D1]-&!@[!$(*#*)&!D1]')
for i, j in smarts.findall(mol):
dihedrals.append(get_dihedral_info(g, (i - 1, j - 1), backbone))
return dihedrals
def get_multiring_atoms_bonds(self, rdk_mol: Mol, smiles):
'''
Not used
'''
atom_ring_times = [0] * rdk_mol.GetNumAtoms()
bond_ring_times = [0] * rdk_mol.GetNumBonds()
# TODO GetRingInfo gives SymmetricSSSR, not TRUE SSSR
ri = rdk_mol.GetRingInfo()
for id_atoms in ri.AtomRings():
for ida in id_atoms:
atom_ring_times[ida] += 1
for id_bonds in ri.BondRings():
for idb in id_bonds:
bond_ring_times[idb] += 1
n_atoms_multiring = len(list(filter(lambda x: x > 1, atom_ring_times)))
n_bonds_multiring = len(list(filter(lambda x: x > 1, bond_ring_times)))
py_mol = pybel.readstring('smi', smiles)
if ri.NumRings() != len(py_mol.sssr):
print(
'WARNING: SymmetricSSSR not equal to TRUE SSSR in rdkit. Use Openbabel instead:',
smiles)
n_atoms_multiring = pybel.Smarts('[R2]').findall(py_mol).__len__()
n_bonds_multiring = n_atoms_multiring - 1
return n_atoms_multiring, n_bonds_multiring
def smarts(s):
if not isinstance(s, bytes):
s = s.encode('ascii')
try:
return pybel.Smarts(s)
except IOError as e:
# Convert pybel's IOError (?!) into a ValueError
raise ValueError(str(e))
def _update(self):
if self.showFragments and self.fragmentSmiles:
loop = self.__update_items(self._items, self._widgets)
elif self.colorFragments and self.selectedFragment:
pattern = pybel.Smarts(self.selectedFragment)
loop = self.__update_items(self._items, self._widgets, pattern)
else:
loop = self.__update_items(self._items, self._widgets)
self.__schedule(loop)
def db_select_molecules(cls=None, subcls=None, type=None, subtype=None, tags=[], substructure="") -> pd.DataFrame:
"""Get a summary frame of molecules in the database
:param tags: a list of tags of the db records (if multiple an 'OR' is taken)
:type tags: list
:param substructure: substructure SMARTS string
:type substructure: str
:return: pandas.core.frame.DataFrame
"""
db = db_connect()
tags_coll = db['tags']
mols_coll = db['molecules']
feats_coll = db['qchem_descriptors']
tags_cur = tags_coll.find({'tag': {'$in': tags}} if tags else {})
tags_df = pd.DataFrame(tags_cur)
filter = {}
if cls != "" and cls is not None:
filter['metadata.class'] = cls
if subcls != "" and subcls is not None:
filter['metadata.subclass'] = subcls
if type != "" and type is not None:
filter['metadata.type'] = type
if subtype != "" and subtype is not None:
filter['metadata.subtype'] = subtype
filter['_id'] = {'$in': tags_df.molecule_id.tolist()}
mols_cur = mols_coll.find(filter)
mols_df = pd.DataFrame(mols_cur)
if 'name' not in mols_df.columns:
mols_df['name'] = None
if substructure:
pattern = pybel.Smarts(substructure)
mols_df['pybel_mol'] = mols_df['can'].map(lambda can: pybel.readstring("smi", can))
mols_df = mols_df[mols_df['pybel_mol'].map(lambda mol: bool(pattern.findall(mol)))]
mols_df = mols_df.drop('pybel_mol', axis=1)
# merge tags in an outer way
df = pd.merge(mols_df, tags_df, how='outer', left_on='_id', right_on='molecule_id', suffixes=('', '_tag'))
# make tags into a list of tags
df['metadata_str'] = df['metadata'].map(repr)
grouped = df.groupby(['can', 'metadata_str'])
# groupby tags
df = pd.concat([grouped['metadata', 'molecule_id', 'name'].first(),
grouped['tag'].apply(list)], axis=1).reset_index().drop('metadata_str', axis=1)
# fetch ids
df['_ids'] = df['molecule_id'].map(lambda mid: [item['_id'] for item in feats_coll.find(
{'molecule_id': ObjectId(mid)}, {'_id': 1})
])
df['num_conformers'] = df['_ids'].map(len)
return df
def index(self, smiles):
double_double = pybel.Smarts('*=**=*')
double_triple = pybel.Smarts('*=**#*')
double_tert = pybel.Smarts('*=*[CX4;H0]')
triple_tert = pybel.Smarts('*#*[CX4;H0]')
r7wired = pybel.Smarts('C1=CC=CC=CC1')
py_mol = pybel.readstring('smi', smiles)
index = [
len(double_double.findall(py_mol)),
len(double_triple.findall(py_mol)),
len(double_tert.findall(py_mol)),
len(triple_tert.findall(py_mol)),
len(r7wired.findall(py_mol)),
]
return np.array(index)
def rearrange_smiles(aa_smiles):
'''Rewrite an amino-acid smiles to start with the N-term and end with
the C-term.'''
mol = pybel.readstring('smi', aa_smiles)
n_term_pat = pybel.Smarts('[$(NCC(O)=O)]')
c_term_pat = pybel.Smarts('[$(OC(=O)CN)]')
#Find location of start and end atoms
n_term_idx = n_term_pat.findall(mol)[0][0]
c_term_idx = c_term_pat.findall(mol)[0][0]
#Rewrite smiles N-term first, then C-term
rearranger = openbabel.OBConversion()
rearranger.SetInAndOutFormats('smi', 'smi')
rearranger.AddOption('f', openbabel.OBConversion.OUTOPTIONS,
str(n_term_idx))
rearranger.AddOption('l', openbabel.OBConversion.OUTOPTIONS,
str(c_term_idx))
outmol = openbabel.OBMol()
rearranger.ReadString(outmol, aa_smiles)
return rearranger.WriteString(outmol).strip()
def find_smarts_hb(dict_smarts, smarts_hb, lig_mol):
# find atoms which can form a hydrogen bond and matches smarts
hb_ph = set()
for smart in dict_smarts[smarts_hb]:
sma = pybel.Smarts(smart)
all_coords = sma.findall(lig_mol)
for coords in all_coords:
for coor in coords:
hb_mol = lig_mol.atoms[coor - 1]
hb_ph.add(hb_mol.coords)
return hb_ph
def smarts_search(mollist, smarts):
ret = []
query = pybel.Smarts(smarts)
print query
for mol in mollist.all():
try:
smiles = pybel.readstring("smi", str(mol.SMILES))
if query.findall(smiles):
ret.append(mol)
except:
pass
return ret
def compile_smarts():
SMARTS = [
'[#6+0!$(*~[#7,#8,F]),SH0+0v2,s+0,S^3,Cl+0,Br+0,I+0]', '[a]',
'[!$([#1,#6,F,Cl,Br,I,o,s,nX3,#7v5,#15v5,#16v4,#16v6,*+1,*+2,*+3])]',
'[!$([#6,H0,-,-2,-3]),$([!H0;#7,#8,#9])]', '[r]'
]
__PATTERNS = []
for smarts in SMARTS:
__PATTERNS.append(pybel.Smarts(smarts))
return __PATTERNS
def pybel_neutralise(pybmol):
pattern = pyb.Smarts(
"[+1!h0!$([*]~[-1,-2,-3,-4]),-1!$([*]~[+1,+2,+3,+4])]")
match_list = pattern.findall(pybmol)
for match in match_list:
atom = pybmol.GetAtom(match[0])
charge = atom.GetFormalCharge()
hnum = atom.GetImplicitHCount()
atom.SetFormalCharge(0)
atom.SetImplicitHCount(hnum - charge)
return pybmol
def dict_toxi(self):
mol_list = []
for smi in [self.smiles]:
l = []
mol = pb.readstring("smi", str(smi))
for k, v in d.iteritems():
k_smarts = pb.Smarts(k)
n = len(k_smarts.findall(mol))
if n == 0:
l.append(0)
else:
for each in v:
d_list = []
v_smarts = pb.Smarts(each)
d_list.append(len(v_smarts.findall(mol)))
if n > sum(d_list):
l.append(1)
elif n == sum(d_list):
l.append(0)
mol_list.append(sum(l))
return mol_list
def substructure_embedding(mol, pattern):
obmol, atom_map = (oasa.pybel_bridge.PybelConverter.
oasa_to_pybel_molecule_with_atom_map(mol))
if not isinstance(pattern, pybel.Smarts):
pattern = pybel.Smarts(pattern)
matches = pattern.findall(obmol)
rev_map = dict(map(reversed, atom_map.iteritems()))
return [map(rev_map.get, match) for match in matches]
def is_alkane(py_mol) -> bool:
import pybel
from .formula import Formula
atom_set = set(Formula(py_mol.formula).atomdict.keys())
if atom_set != {'C', 'H'}:
return False
for s in ['[CX2]', '[CX3]', 'c', '[#6;v0,v1,v2,v3]']:
if pybel.Smarts(s).findall(py_mol) != []:
return False
else:
return True
def find_smarts(dict_smarts, smarts, lig_mol):
# find groups of atoms which matches smarts
coords_interaction = []
for smart in dict_smarts[smarts]:
sma = pybel.Smarts(smart)
all_coords = sma.findall(lig_mol)
for coords in all_coords:
c_ph = []
for coor in coords:
c_mol = lig_mol.atoms[coor - 1]
c_ph.append(c_mol.coords)
coords_interaction.append(c_ph)
return coords_interaction
def unique_toxi(self):
mol_list = []
for smi in [self.smiles]:
l = []
mol = pb.readstring("smi", str(smi))
for toxi in all_toxi:
smarts = pb.Smarts(toxi)
if len(smarts.findall(mol)) > 0:
l.append(1)
else:
l.append(0)
mol_list.append(sum(l))
return mol_list
def atoms_nitrophenols(mol, phenol, nitro):
## returns phenols for which nitro groups are found in same ring
def is_part_of_ring(ring, group):
ring_atoms = atom_indices_ring(r, mol) #not the most efficient
return len([idx for idx in group if idx in ring_atoms]) > 0
def atom_indices_ring(ring, mol):
return [a.idx for a in mol.atoms if ring.IsInRing(a.idx)]
def atom_indices_group(groups):
return list(chain.from_iterable(groups))
_phenol = pybel.Smarts(phenol).findall(mol)
_nitro = pybel.Smarts(nitro).findall(mol)
_rings = [ring for ring in mol.sssr if ring.IsAromatic()]
atomlist_ring = [] # list of rings
atomlist_nitrophenol = [] # list of (nitro)phenol groups
for r in _rings:
part = {'phenol': [], 'nitro': []}
for x in _phenol:
if not is_part_of_ring(r, x):
continue
part['phenol'].append(x)
for y in _nitro:
if not is_part_of_ring(r, y):
continue
part['nitro'].append(y)
if part['phenol'] and part['nitro']:
atomlist_ring.append(
set(
atom_indices_ring(r, mol) +
atom_indices_group(part['phenol']) +
atom_indices_group(part['nitro'])))
atomlist_nitrophenol += part['phenol']
# returning of atomlist_ring is optional
# but the phenol groups are what are really counted so this is what is returned
return atomlist_nitrophenol
def MatchPlattsBGroups(self, smiles):
# Load functional group database
current_dir = os.getcwd()
filepath = os.path.join(current_dir, 'groups.xls')
wb = xlrd.open_workbook(filepath)
wb.sheet_names()
data = wb.sheet_by_name(u'PlattsB')
col1 = data.col_values(0)
col2 = data.col_values(1)
col3 = data.col_values(2)
databaseB = []
for (SMART, name, B) in zip(col1, col2, col3):
databaseB.append(functionalgroup(SMART, name, B))
platts_B = 0
mol = pybel.readstring("smi", smiles)
for x in databaseB:
# Initialize with dummy SMLES to check for validity of real one
smarts = pybel.Smarts("CC")
smarts.obsmarts = ob.OBSmartsPattern()
success = smarts.obsmarts.Init(x.smarts.__str__())
if success:
smarts = pybel.Smarts(x.smarts.__str__())
else:
print "Invalid SMARTS pattern", x.smarts.__str__()
break
matched = smarts.findall(mol)
x.num = len(matched)
if (x.num > 0):
print "Found group", x.smarts.__str__(
), 'named', x.name, 'with contribution', x.value, 'to B', x.num, 'times'
platts_B += (x.num) * (x.value)
self.B = platts_B + 0.071
