def remove_labile_protons(sdffile, lbls, shifts):
f = sdffile.split('.sdf')[0] + '.sdf'
obconversion = OBConversion()
obconversion.SetInFormat("sdf")
obmol = OBMol()
obconversion.ReadFile(obmol, f)
CI = []
for atom in OBMolAtomIter(obmol):
if atom.GetAtomicNum() == 1:
for NbrAtom in OBAtomAtomIter(atom):
if (NbrAtom.GetAtomicNum() == 8):
CI.append('H' + str(atom.GetIndex() + 1))
#remove these carbons
for C in CI:
ind = lbls.index(C)
lbls.remove(C)
for l in shifts:
l.pop(ind)
return lbls, shifts
def methyl_protons(file):
obconversion = OBConversion()
obconversion.SetInFormat("sdf")
obmol = OBMol()
obconversion.ReadFile(obmol, file)
methyl_protons = []
for atom in OBMolAtomIter(obmol):
count = 0
nbrprotons = []
for NbrAtom in OBAtomAtomIter(atom):
if (atom.GetAtomicNum() == 6) & (NbrAtom.GetAtomicNum() == 1):
l = NbrAtom.GetIndex()
count += 1
nbrprotons.append('H' + str(l + 1))
if count == 3:
methyl_protons.append(nbrprotons)
return methyl_protons
def convert_babel_to_system(mol):
"""
Convert a BigDFT fragment to an open babel molecule.
Args:
mol (openbabel.OBMol): the molecule to convert.
Returns:
(BigDFT.Systems.System): bigdft system.
"""
from BigDFT.IO import read_mol2
from openbabel.openbabel import OBConversion
# py2 workaround
from sys import version_info
if version_info[0] < 3:
from io import BytesIO as StringIO
else:
try:
from io import StringIO
except ImportError:
from StringIO import StringIO
conv = OBConversion()
conv.SetOutFormat("mol2")
sval = StringIO(conv.WriteString(mol))
return read_mol2(sval)
def convert_system_to_babel(sys):
"""
Convert a BigDFT system to an open babel molecule.
Args:
sys (BigDFT.Systems.System): the system to convert.
Returns:
(openbabel.OBMol): an open babel type molecule.
"""
from BigDFT.IO import write_pdb
from openbabel.openbabel import OBMol, OBConversion
# py2 workaround
from sys import version_info
if version_info[0] < 3:
from io import BytesIO as StringIO
else:
try:
from io import StringIO
except ImportError:
from StringIO import StringIO
# We convert by way of pdb file.
conv = OBConversion()
conv.SetInFormat("pdb")
sval = StringIO()
write_pdb(sys, sval)
mol = OBMol()
conv.ReadString(mol, sval.getvalue())
return mol
def Karplus(f, inputformat):
obconversion = OBConversion()
obconversion.SetInFormat(inputformat)
obmol = OBMol()
obconversion.ReadFile(obmol, f)
obmol.ConnectTheDots()
obmol.Kekulize()
DihedralHs = []
for atom in OBMolAtomIter(obmol):
if atom.GetAtomicNum() == 1:
DihedNeighbours = GetDihedralHs(atom)
if DihedNeighbours != 0:
DihedralHs.append([atom.GetIdx()] + DihedNeighbours)
if len(DihedralHs) == 0:
print("No dihedral protons found, Karplus J value prediction " + \
"impossible, quitting.")
quit()
Jmatrix, Jlabels = CalcJMatrix(obmol, DihedralHs)
Jlabels = [str(x) for x in Jlabels]
return Jmatrix, Jlabels
def RestoreNumsSDF(f, fold, AuxInfo):
#Read molecule from file
obconversion = OBConversion()
obconversion.SetInFormat("sdf")
obmol = OBMol()
obconversion.ReadFile(obmol, f)
#Get the atoms Hs are connected to
oldHcons = GetHcons(fold)
#translate the H connected atoms to the new numbering system
amap = GetInchiRenumMap(AuxInfo)
for i in range(0, len(oldHcons)):
oldHcons[i][1] = amap.index(oldHcons[i][1]) + 1
newHcons = []
temp = []
i = 0
for atom in OBMolAtomIter(obmol):
idx = atom.GetIdx()
anum = atom.GetAtomicNum()
#If atom is hydrogen, check what it is connected to
if anum == 1:
for NbrAtom in OBAtomAtomIter(atom):
newHcons.append([idx, NbrAtom.GetIdx()])
#Pick the temporary atom
temp.append(atom)
for i in range(0, len(newHcons)):
conatom = newHcons[i][1]
for b in range(0, len(oldHcons)):
if conatom == oldHcons[b][1]:
amap.append(oldHcons[b][0])
#remove the number, so that it doesn't get added twice
oldHcons[b][1] = 0
newmol = OBMol()
added = []
for i in range(1, len(amap) + 1):
newn = amap.index(i)
newmol.AddAtom(temp[newn])
added.append(newn)
#Final runthrough to check that all atoms have been added,
#tautomeric protons can be missed. If tautomeric proton tracking
#is implemented this can be removed
for i in range(0, len(temp)):
if not i in added:
newmol.AddAtom(temp[i])
#Restore the bonds
newmol.ConnectTheDots()
newmol.PerceiveBondOrders()
#Write renumbered molecule to file
obconversion.SetOutFormat("sdf")
obconversion.WriteFile(newmol, f)
def GetHcons(f):
obconversion = OBConversion()
obconversion.SetInFormat("sdf")
obmol = OBMol()
obconversion.ReadFile(obmol, f)
Hcons = []
for atom in OBMolAtomIter(obmol):
idx = atom.GetIdx()
anum = atom.GetAtomicNum()
if anum == 1:
for NbrAtom in OBAtomAtomIter(atom):
Hcons.append([idx, NbrAtom.GetIdx()])
return Hcons
def labile_protons(file):
obconversion = OBConversion()
obconversion.SetInFormat("sdf")
obmol = OBMol()
obconversion.ReadFile(obmol, file)
count = 0
for atom in OBMolAtomIter(obmol):
for NbrAtom in OBAtomAtomIter(atom):
if (atom.GetAtomicNum() == 8) & (NbrAtom.GetAtomicNum() == 1):
count += 1
return count
def runvina(infile, outfile, receptor, tmp_file='test.pdbqt', vina=None):
obconversion = OBConversion()
obconversion.SetInFormat("sdf")
obconversion.SetOutFormat("pdbqt")
obmol = OBMol()
notatend = obconversion.ReadFile(obmol, infile)
obmol2 = OBMol(obmol)
ofs = pybel.Outputfile("sdf", outfile, overwrite=True)
pbar = tqdm()
while notatend:
pbar.update(1)
if obconversion.WriteFile(obmol, tmp_file):
try:
x = subprocess.check_output([
vina, "--score_only", "--receptor", receptor, "--ligand",
tmp_file
],
shell=False)
# x2 = subprocess.check_output(["/Users/austin/Downloads/rf-score-4/rf-score", "/Users/austin/Downloads/rf-score-4/pdbbind-2014-refined.rf", receptor, tmp_file])
# print(x2)
mol2 = pybel.Molecule(obmol2)
mol2.data.update({'AutodockVinaRescoreOnly': str(get_aff(x))})
ofs.write(mol2)
except subprocess.CalledProcessError as e:
print(e)
ofs.write(obmol)
except ValueError as e:
print(e)
ofs.write(obmol)
else:
print("error writing")
obmol = OBMol()
notatend = obconversion.Read(obmol)
obmol2 = OBMol(obmol)
pbar.close()
print("FAILED")
def FixTautProtons(f, inchi, AuxInfo):
#Get tautomeric protons and atoms they are connected to from Inchi
TautProts = GetTautProtons(inchi)
amap = GetInchiRenumMap(AuxInfo)
#get the correspondence of the Inchi numbers to the source numbers
hmap = []
for taut in TautProts:
for heavyatom in range(1, len(taut)):
hmap.append([int(taut[heavyatom]), amap[int(taut[heavyatom]) - 1]])
#Read molecule from file
obconversion = OBConversion()
obconversion.SetInFormat("sdf")
obmol = OBMol()
obconversion.ReadFile(obmol, f)
Fixprotpos = []
for heavyatom in hmap:
atom = obmol.GetAtom(heavyatom[1])
for nbratom in OBAtomAtomIter(atom):
if nbratom.GetAtomicNum() == 1:
Fixprotpos.append(heavyatom[0])
draftFH = []
for i in range(0, len(Fixprotpos)):
if Fixprotpos[i] not in [a[0] for a in draftFH]:
draftFH.append([Fixprotpos[i], Fixprotpos.count(Fixprotpos[i])])
fixedlayer = '/f/h'
for h in draftFH:
if h[1] == 1:
fixedlayer = fixedlayer + str(h[0]) + 'H,'
else:
fixedlayer = fixedlayer + str(h[0]) + 'H' + str(h[1]) + ','
resinchi = inchi + fixedlayer[:-1]
return resinchi
def compute_smiles(sys):
"""
Computes the SMILES representation of a given system.
Args:
sys (BigDFT.System.Systems): the system to compute the
representation of.
Return:
(str): the smiles representation of this molecule.
"""
from openbabel.openbabel import OBConversion
conv = OBConversion()
mol = convert_system_to_babel(sys)
conv.SetOutFormat("SMI")
retstr = conv.WriteString(mol)
retstr = retstr.replace("\n", "")
retstr = retstr.replace("\t", "")
return retstr
def mol_fragments(mole,outfile):
obconv = OBConversion()
obconv.SetOutFormat("xyz")
c = 1
for atom, exp,dft,diff in zip(OBMolAtomIter(mole)):
# if this is a carbon atom start a breadth first search for other carbon atoms with depth specified
# create a new mol instance
new_mol = OBMol()
# add this atom
# new_mol.AddAtom(atom)
fragment_ind = []
l = atom.GetIndex()
fragment_ind.append(l)
# for iteration depth radius
old_queue = [atom]
for iteration in range(0, 3):
new_queue = []
for a in old_queue:
for atom2 in OBAtomAtomIter(a):
i = atom2.GetIndex()
# if the atom has not been seen before add it to the fragment ind list and to the new molecule
if i not in fragment_ind:
new_queue.append(atom2)
fragment_ind.append(i)
# new_mol.AddAtom(atom2)
old_queue = copy.copy(new_queue)
fragment_ind = [fragment_ind[0]] + sorted(fragment_ind[1:])
for i in fragment_ind:
for a in OBMolAtomIter(mole):
if a.GetIndex() == i:
new_mol.AddAtom(a)
f = open(outfile + "frag" + str(l).zfill(3) + ".xyz", "w+")
f.write(str(new_mol.NumAtoms()) + "\n\n")
i = 0
for atom in OBMolAtomIter(new_mol):
f.write(atom.GetType()[0] + " " + str(atom.GetX()) + " " + str(atom.GetY()) + " " + str(
atom.GetZ()) + "\n")
i+=1
f.close()
c += 1
def ProcessIsomers(dp5Data, Isomers,Settings):
OutputFolder = Path(Settings.OutputFolder)
# extract calculated and experimental shifts and add to dp5Data instance
# Carbon
# make sure any shifts with missing peaks are removed from all isomers
removedC = []
no_exp_data = 0
for iso in Isomers:
if len(iso.Cexp) == 0:
no_exp_data +=1
if no_exp_data == len(Isomers):
print("no experimental NMR data provided... quitting")
quit()
for iso in Isomers:
dp5Data.Cexp.append([])
dp5Data.Cshifts.append([])
dp5Data.Clabels.append([])
dp5Data.Cinds.append([])
dp5Data.ConfCshifts.append([])
a_ind = 0
exp_inds = []
for shift, exp, label in zip(iso.Cshifts, iso.Cexp, iso.Clabels):
if exp != '':
dp5Data.Cshifts[-1].append(shift)
dp5Data.Cexp[-1].append(exp)
dp5Data.Clabels[-1].append(label)
dp5Data.Cinds[-1].append(int(label[1:]) - 1)
exp_inds.append(a_ind)
a_ind += 1
if len(iso.ConformerCShifts) > 0:
for conf_shifts in iso.ConformerCShifts:
dp5Data.ConfCshifts[-1].append( [conf_shifts[e] for e in exp_inds])
#write qml compound objects and atomic representations
#check the number of atoms in the structures
#if there are less than 86 (max number of atoms in a molecule in the training set) atoms
if dp5Data.Atom_number < 86:
for iso in Isomers:
#open new xyz file
InputFile = Path(iso.InputFile)
#find conformer with the lowest energy
dp5Data.AtomReps.append([])
for i,geom in enumerate(iso.DFTConformers):
xyz_file = open(str(OutputFolder / "dp5" /InputFile.stem) + "_" +str(i).zfill(3) + ".xyz", "w")
xyz_file.write(str(len(iso.Atoms)) + "\n" + "\n")
for atom, coords in zip(iso.Atoms, geom):
xyz_file.write(atom + " " + str(coords[0]) + " " + str(coords[1]) + " " + str(coords[2]) + "\n")
xyz_file.close()
dp5Data.Compounds.append(qml.Compound(xyz = str(Settings.OutputFolder/"dp5"/ InputFile.stem) +"_"+ str(i).zfill(3) + ".xyz"))
dp5Data.Compounds[-1].generate_fchl_representation(max_size=86, cut_distance=c_distance)
dp5Data.AtomReps[-1].append([])
for C_l in iso.Clabels:
ind = int(C_l.split("C")[1])
dp5Data.AtomReps[-1][-1].append(dp5Data.Compounds[-1].representation[ind])
#otherwise we need to fragment the molecule to radius of 3
else:
for iso in Isomers:
#open new xyz file
InputFile = Path(iso.InputFile)
#find conformer with the lowest energy
dp5Data.AtomReps.append([])
for i,geom in enumerate(iso.DFTConformers):
xyz_file = open(str(OutputFolder / "dp5" /InputFile.stem) + "_" +str(i).zfill(3) + ".xyz", "w")
xyz_file.write(str(len(iso.Atoms)) + "\n" + "\n")
for atom, coords in zip(iso.Atoms, geom):
xyz_file.write(atom + " " + str(coords[0]) + " " + str(coords[1]) + " " + str(coords[2]) + "\n")
xyz_file.close()
#now need to fragment the molecule and generate these representations
#build ob mol
obconversion = OBConversion()
obconversion.SetInFormat("sdf")
m = OBMol()
obconversion.ReadFile(m, iso.InputFile)
os.mkdir(str(OutputFolder / "dp5" /InputFile.stem) + "_" +str(i).zfill(3) + "_fragments")
mol_fragments(m,str(OutputFolder / "dp5" /InputFile.stem) + "_" +str(i).zfill(3) + "_fragments")
conf_rep = []
for xyz_frag in sorted(os.listdir( str(OutputFolder / "dp5" /InputFile.stem) + "_" +str(i).zfill(3) + "_fragments")):
c = qml.Compound(xyz=str(OutputFolder / "dp5" /InputFile.stem) + "_" +str(i).zfill(3) + "_fragments/" + xyz_frag)
c.generate_fchl_representation(max_size=54, cut_distance=c_distance)
conf_rep.append(c.representation[0])
dp5Data.AtomReps[-1].append([])
for C_l in iso.Clabels:
ind = int(C_l.split("C")[1])
dp5Data.AtomReps[-1][-1].append(conf_rep[ind])
return dp5Data
def main(f, settings):
"""
Find the axis atoms
Find all the atoms to be rotated
Rotate it and the substituents to the other side of the plane
"""
obconversion = OBConversion()
obconversion.SetInFormat("sdf")
obmol = OBMol()
obconversion.ReadFile(obmol, f)
obmol.ConnectTheDots()
#Find the atoms composing furan ring
Rings = obmol.GetSSSR()
furan = []
for ring in Rings:
if len(settings.RingAtoms) == 5:
if all(x in ring._path for x in settings.RingAtoms):
furan = ring
break
else:
if ring.Size() == 5 and not ring.IsAromatic():
furan = ring
break
if furan == []:
"No five membered rings to rotate. Quitting..."
quit()
#Find the plane of the 5-membered ring and the outlying atom
norm, d, outAtom = FindFuranPlane(obmol, furan)
#Find the atoms connected to the outlying atom and sort them
#as either part of the ring(axis atoms) or as atoms to be rotated
AxisAtoms = []
RotAtoms = []
for NbrAtom in OBAtomAtomIter(outAtom):
#if NbrAtom.IsInRingSize(5):
if furan.IsInRing(NbrAtom.GetIdx()):
AxisAtoms.append(NbrAtom)
else:
RotAtoms.append(NbrAtom)
FindSubstAtoms(NbrAtom, outAtom, RotAtoms)
#Simple switch to help detect if the atoms are rotated the right way
WasAbove90 = False
angle = FindRotAngle(AxisAtoms[0], AxisAtoms[1], outAtom, norm)
if angle > 0.5 * pi:
WasAbove90 = True
rotangle = 2 * (angle - 0.5 * pi)
else:
WasAbove90 = False
rotangle = 2 * (0.5 * pi - angle)
OldAtomCoords = outAtom.GetVector()
print("Atom " + str(outAtom.GetAtomicNum()) + " will be rotated by " +\
str(rotangle*57.3) + ' degrees')
RotateAtom(outAtom, AxisAtoms[0], AxisAtoms[1], rotangle)
angle2 = FindRotAngle(AxisAtoms[0], AxisAtoms[1], outAtom, norm)
#if the atom is on the same side of the plane as it was,
# it has been rotated in the wrong direction
if ((angle2 > 0.5 * pi) and WasAbove90) or ((angle2 < 0.5 * pi)
and not WasAbove90):
#Flip the sign of the rotation angle, restore the coords
#and rotate the atom in the opposite direction
print("Atom was rotated the wrong way, switching the direction")
rotangle = -rotangle
outAtom.SetVector(OldAtomCoords)
RotateAtom(outAtom, AxisAtoms[0], AxisAtoms[1], rotangle)
RotatedAtoms = [] # Index to make sure that atoms are not rotated twice
for atom in RotAtoms:
if atom not in RotatedAtoms:
RotateAtom(atom, AxisAtoms[0], AxisAtoms[1], rotangle)
RotatedAtoms.append(atom)
else:
print("Atom already rotated, skipping")
obconversion.SetOutFormat("sdf")
obconversion.WriteFile(obmol, f[:-4] + 'rot.sdf')