def otherinteractions(resdict, residuechoice, protein, ligand,
backboneatomlist):
for residue in OBResidueIter(protein):
residuename = residue.GetName()
if residuename in residuechoice:
for atom in OBResidueAtomIter(residue):
if not atom.IsHydrogen():
for atomlig in OBMolAtomIter(ligand):
if not atomlig.IsHydrogen():
distance = atom.GetDistance(atomlig)
if distance <= 4.5:
if isnonpolar(atomlig) & isnonpolar(atom):
#print "POLAR"
resdict[residuename] |= bitarray('1000000')
if distance <= 4.0:
setformalcharge(atom, backboneatomlist)
fakebbatomlist = []
setformalcharge(atomlig, fakebbatomlist)
if (atom.GetFormalCharge() > 0) & (
atomlig.GetFormalCharge() < 0):
#print "NEGATIVE"
resdict[residuename] |= bitarray(
'0000010')
if (atom.GetFormalCharge() < 0) & (
atomlig.GetFormalCharge() > 0):
#print "POSITIVE"
resdict[residuename] |= bitarray(
'0000001')
if distance <= 3.5:
if atom.IsHbondDonor(
) & atomlig.IsHbondAcceptor():
for neighborDon in OBAtomAtomIter(
atom):
if neighborDon.IsHydrogen():
angle = atom.GetAngle(
neighborDon, atomlig)
if angle > 135.0:
resdict[
residuename] |= bitarray(
'0001000')
# donorresidue = atom.GetResidue()
# for atomres in OBResidueAtomIter(donorresidue):
# if atomres.IsHydrogen() & atomres.IsConnected(atom):
# angle = atom.GetAngle(atomres, atomlig)
# if angle>135.0:
# resdict[residuename] |= bitarray('0001000')
if atom.IsHbondAcceptor(
) & atomlig.IsHbondDonor():
for neighborDon in OBAtomAtomIter(
atomlig):
if neighborDon.IsHydrogen():
#print True
angle = atomlig.GetAngle(
neighborDon, atom)
#print angle
if angle > 135.0:
#print "ACCEPTOR"
resdict[
residuename] |= bitarray(
'0000100')
def hbonddockprot(resdict, residuechoice, protein, ligand):
for residue in OBResidueIter(protein):
residuename = residue.GetName()
if residuename in residuechoice:
for atom in OBResidueAtomIter(residue):
if not atom.IsHydrogen():
for atomlig in OBMolAtomIter(ligand):
if not atomlig.IsHydrogen():
distance = atom.GetDistance(atomlig)
if distance <= 3.5:
if atom.IsHbondDonor(
) & atomlig.IsHbondAcceptor():
for neighborDon in OBAtomAtomIter(atom):
if neighborDon.IsHydrogen():
angle = atom.GetAngle(
neighborDon, atomlig)
if angle > 135.0:
resdict[
residuename] |= bitarray(
'0001000')
if atom.IsHbondAcceptor(
) & atomlig.IsHbondDonor():
for neighborDon in OBAtomAtomIter(atomlig):
if neighborDon.IsHydrogen():
angle = atomlig.GetAngle(
neighborDon, atom)
if angle > 135.0:
resdict[
residuename] |= bitarray(
'0000100')
def hydrogen_nbrs(atom):
'''
Returns an iterator over all hydrogens of an atom. Required to calculate hydrogen
bonds.
'''
for nbr in OBAtomAtomIter(atom):
if nbr.IsHydrogen():
yield nbr
def neighbors(self):
return [Atom(a) for a in OBAtomAtomIter(self.OBAtom)]
def otherinteractions2(resdict, residuechoice, protein, ligand,
backboneatomlist):
RigidHdonors = ['HIS', 'GLN', 'ASN', 'ARG', 'TRP']
for residue in OBResidueIter(protein):
residuename = residue.GetName()
if residuename in residuechoice:
for atom in OBResidueAtomIter(residue):
if not atom.IsHydrogen():
for atomlig in OBMolAtomIter(ligand):
print atomlig
if not atomlig.IsHydrogen():
distance = atom.GetDistance(atomlig)
if distance <= 4.5:
if isnonpolar(atomlig) & isnonpolar(atom):
resdict[residuename] |= bitarray('1000000')
if distance <= 4.0:
setformalcharge(atom, backboneatomlist)
fakebbatomlist = []
setformalcharge(atomlig, fakebbatomlist)
if (atom.GetFormalCharge() > 0) & (
atomlig.GetFormalCharge() < 0):
resdict[residuename] |= bitarray(
'0000010')
if (atom.GetFormalCharge() < 0) & (
atomlig.GetFormalCharge() > 0):
resdict[residuename] |= bitarray(
'0000001')
if distance <= 3.5:
if residuename[:3] in RigidHdonors:
if atom.IsHbondDonor(
) & atomlig.IsHbondAcceptor():
for neighborDon in OBAtomAtomIter(
atom):
if neighborDon.IsHydrogen(
):
angle = atom.GetAngle(
neighborDon,
atomlig)
if angle > 135.0:
resdict[
residuename] |= bitarray(
'0001000')
elif atom.GetIdx() in backboneatomlist:
if atom.IsHbondDonor(
) & atomlig.IsHbondAcceptor():
for neighborDon in OBAtomAtomIter(
atom):
if neighborDon.IsHydrogen(
):
angle = atom.GetAngle(
neighborDon,
atomlig)
if angle > 135.0:
resdict[
residuename] |= bitarray(
'0001000')
if atom.IsHbondAcceptor(
) & atomlig.IsHbondDonor():
for neighborDon in OBAtomAtomIter(
atomlig):
if neighborDon.IsHydrogen():
angle = atomlig.GetAngle(
neighborDon, atom)
if angle > 135.0:
resdict[
residuename] |= bitarray(
'0000100')
def otherinteractions(l_bit, l_res_choices, st_protein, st_lig,
l_serial_atom_backbone):
for residue in OBResidueIter(st_protein):
residuename = residue.GetName()
if residuename in l_res_choices:
for atom in OBResidueAtomIter(residue):
if not atom.IsHydrogen():
for atomlig in OBMolAtomIter(st_lig):
if not atomlig.IsHydrogen():
distance = atom.GetDistance(atomlig)
if distance <= 4.5:
if isnonpolar(atomlig) & isnonpolar(atom):
l_bit[residuename] |= bitarray('1000000')
if distance <= 4.0:
setformalcharge(atom,
l_serial_atom_backbone)
fakebbatomlist = []
setformalcharge(atomlig, fakebbatomlist)
if (atom.GetFormalCharge() > 0) & (
atomlig.GetFormalCharge() < 0):
l_bit[residuename] |= bitarray(
'0000010')
if (atom.GetFormalCharge() < 0) & (
atomlig.GetFormalCharge() > 0):
l_bit[residuename] |= bitarray(
'0000001')
if distance <= 3.5:
if atom.IsHbondDonor(
) & atomlig.IsHbondAcceptor():
for neighborDon in OBAtomAtomIter(
atom):
if neighborDon.IsHydrogen():
for neighborAcc in OBAtomAtomIter(
atomlig):
angle = neighborAcc.GetAngle(
atomlig,
neighborDon)
if angle > 135.0:
l_bit[
residuename] |= bitarray(
'0001000')
# donorresidue = atom.GetResidue()
# for atomres in OBResidueAtomIter(donorresidue):
# if atomres.IsHydrogen() & atomres.IsConnected(atom):
# angle = atom.GetAngle(atomres, atomlig)
# if angle>135.0:
# l_bit[residuename] |= bitarray('0001000')
if atom.IsHbondAcceptor(
) & atomlig.IsHbondDonor():
for neighborDon in OBAtomAtomIter(
atomlig):
if neighborDon.IsHydrogen():
for neighborAcc in OBAtomAtomIter(
atom):
angle = neighborAcc.GetAngle(
atom, neighborDon)
if angle > 135.0:
l_bit[
residuename] |= bitarray(
'0000100')
def load_data(self,scaffoldName,options={}):
'''
This method loads the scaffold data from a cIdx file <scaffold name>.cIdx and scaffold name>.sdf
Args:
string scaffoldName: the name of the scaffold
Raises:
IOError
ScaffoldNullPopularity
ScaffoldCorrupted
Example:
load_data("s000000001")
'''
verbose =options['verbose'] if options.has_key('verbose') else False
debug =options['debug'] if options.has_key('debug') else False
#------------ data directly loaded from NPSD without further processing ------------------------------
cIdxf=open(scaffoldName+".cIdx")
# Processing the cIdx file
Idx=0
configNumber=0
sidechainBlock=0
positionsNumber=0
position=-1
for line in cIdxf:
if Idx==0:
a=line.strip().split()
self.name = a[0]
self.popularity = int(a[1])
if not self.popularity:
break
Idx+=1
continue
if Idx==1:
a=line.strip().split()
self.max_possible_weight = float(a[0])
self.min_possible_weight = float(a[1])
Idx+=1
continue
if Idx==2:
configNumber=float(line.strip().split()[0])
sidechainBlock=3+configNumber
Idx+=1
continue
if Idx>=3 and Idx<sidechainBlock:
a=line.strip().split()
nonHydrogenPositions=int(a[0])
self.configurations.append([int(x) for x in a[1:nonHydrogenPositions+1]])
Idx+=1
continue
if Idx>=3 and Idx==sidechainBlock:
self.scaffoldLength=int(line.split()[0])
for i in range(0,self.scaffoldLength):
self.sidechainlist.append([])
Idx+=1
continue
if Idx>=3 and Idx>sidechainBlock:
a=line.strip().split('\t')
if len(a)==2: #new position
position =int(a[0])
sidechainsNumber =int(a[1])
else:
self.sidechainlist[position].append([float(a[4]),float(a[3]),a[0]])
Idx+=1
continue
Idx+=1
if position+1<self.scaffoldLength:
raise ScaffoldCorrupted("Warning: some positions are missing: %d/%d. The file is probably corrupted." %(position+1,self.scaffoldLength))
if not self.popularity:
cIdxf.close()
raise ScaffoldNullPopularity("Warning: the popularity of this scaffold is 0.")
# Loading the scaffold molecule (.sdf file) thanks to pybel
self.structure = pybel.readfile("sdf",scaffoldName+".sdf").next()
self.scaffold_weight_plus_hydrogens = self.structure.molwt
#------------ data obtained by processing -----------------------------------------------------------
c=0
for conf in self.configurations:
# we add hydrogens to be able to count them. Nevertheless it doesn't affect the molecular weight which always includes the hydrogens
self.structure.OBMol.AddHydrogens()
# \we compute the total number of hydrogens
total_number_of_hydrogenes =0
for atom in [a for a in self.structure if not a.OBAtom.GetAtomicNum()==1]:
if atom.idx-1 in conf:
number_of_hydrogenes =len([x for x in OBAtomAtomIter(atom.OBAtom) if x.GetAtomicNum()==1])
total_number_of_hydrogenes +=number_of_hydrogenes
#
# \Dealing with the mass
# we remove the hydrogens because we don't need them anymore
self.structure.OBMol.DeleteHydrogens()
# we compute the total mass of hydrogens at non hydrogen side chain positions
total_hydrogene_mass = total_number_of_hydrogenes * 1.00794
# we compute the mass of the scaffold without the non hydrogen positions
hs_weight =self.scaffold_weight_plus_hydrogens - total_hydrogene_mass
self.scaffold_weight_rel2configs.append(hs_weight)
#
# \Dealing with the probability
# the condition a.atomicnum>0 is to ensure that we don't count the "atoms" R# that are just used to number the positions on the scaffold
hs_probabilities=[1]
for atom in [ a for a in self.structure if not a.idx-1 in conf and a.atomicnum>0] :
idx =atom.idx-1
# \now we find the hydrogen sidechain in the list and save its probability
for sidechain in self.sidechainlist[idx]:
smile =sidechain[2]
probability =sidechain[1]
original_smile = smile
# if not H, that SMILE cannot be an hydrogen sidechain
if 'H' in smile:
for char in smiles_char:
#removing all reserved characters from the SMILE. Only atom names remain
smile =smile.replace(char,'')
# \now we check that only H are present in the string, otherwise it is not a hydrogen side chain
ishydrogen = True
for char in smile:
if not char=='H':
ishydrogen=False
#
if ishydrogen:
hs_probabilities.append(probability)
#
#//end for atom in [ a for a in self.structure if not a.idx-1 in conf and a.atomicnum>0]
hs_cumulative_probability = reduce(mul,hs_probabilities)
self.scaffold_probability_rel2configs.append(hs_cumulative_probability)
c+=1
#//end for conf in self.configurations
cIdxf.close()
self.number_of_compounds=self.count_compounds()
def pication(rings, pos_charged, protcharged):
"""Return all pi-Cation interaction between aromatic rings and positively charged groups.
For tertiary and quaternary amines, check also the angle between the ring and the nitrogen.
"""
data = namedtuple(
'pication',
'ring charge distance offset type restype resnr reschain restype_l resnr_l reschain_l protcharged'
)
pairings = []
if len(rings) == 0 or len(pos_charged) == 0:
return pairings
for ring in rings:
c = ring.center
for p in pos_charged:
d = euclidean3d(c, p.center)
# Project the center of charge into the ring and measure distance to ring center
proj = projection(ring.normal, ring.center, p.center)
offset = euclidean3d(proj, ring.center)
if not config.MIN_DIST < d < config.PICATION_DIST_MAX or not offset < config.PISTACK_OFFSET_MAX:
continue
if type(p).__name__ == 'lcharge' and p.fgroup == 'tertamine':
# Special case here if the ligand has a tertiary amine, check an additional angle
# Otherwise, we might have have a pi-cation interaction 'through' the ligand
n_atoms = [
a_neighbor
for a_neighbor in OBAtomAtomIter(p.atoms[0].OBAtom)
]
n_atoms_coords = [(a.x(), a.y(), a.z()) for a in n_atoms]
amine_normal = np.cross(
vector(n_atoms_coords[0], n_atoms_coords[1]),
vector(n_atoms_coords[2], n_atoms_coords[0]))
b = vecangle(ring.normal, amine_normal)
# Smallest of two angles, depending on direction of normal
a = min(b, 180 - b if not 180 - b < 0 else b)
if not a > 30.0:
resnr, restype = whichresnumber(
ring.atoms[0]), whichrestype(ring.atoms[0])
reschain = whichchain(ring.atoms[0])
resnr_l, restype_l = whichresnumber(
p.orig_atoms[0]), whichrestype(p.orig_atoms[0])
reschain_l = whichchain(p.orig_atoms[0])
contact = data(ring=ring,
charge=p,
distance=d,
offset=offset,
type='regular',
restype=restype,
resnr=resnr,
reschain=reschain,
restype_l=restype_l,
resnr_l=resnr_l,
reschain_l=reschain_l,
protcharged=protcharged)
pairings.append(contact)
break
resnr = whichresnumber(
p.atoms[0]) if protcharged else whichresnumber(ring.atoms[0])
resnr_l = whichresnumber(
ring.orig_atoms[0]) if protcharged else whichresnumber(
p.orig_atoms[0])
restype = whichrestype(
p.atoms[0]) if protcharged else whichrestype(ring.atoms[0])
restype_l = whichrestype(
ring.orig_atoms[0]) if protcharged else whichrestype(
p.orig_atoms[0])
reschain = whichchain(p.atoms[0]) if protcharged else whichchain(
ring.atoms[0])
reschain_l = whichchain(
ring.orig_atoms[0]) if protcharged else whichchain(
p.orig_atoms[0])
contact = data(ring=ring,
charge=p,
distance=d,
offset=offset,
type='regular',
restype=restype,
resnr=resnr,
reschain=reschain,
restype_l=restype_l,
resnr_l=resnr_l,
reschain_l=reschain_l,
protcharged=protcharged)
pairings.append(contact)
return filter_contacts(pairings)
for residue in protein.residues:
# if residue.OBResidue.GetAminoAcidProperty(8): # get hydrophobic aa
for atom in residue:
# get
if residue.OBResidue.GetAminoAcidProperty(
8) and atom.atomicnum == 6:
protein_hydrophobe.append(atom.coords)
protein_hydrophobe_res.append(
residue.name[:3] + str(residue.OBResidue.GetNum()))
if atom.OBAtom.IsHbondDonor():
protein_donors.append(atom.coords)
protein_donors_res.append(residue.name[:3] +
str(residue.OBResidue.GetNum()))
nbr = []
for nbr_atom in [
pybel.Atom(x) for x in OBAtomAtomIter(atom.OBAtom)
]:
if not nbr_atom.OBAtom.IsHbondDonorH():
nbr.append(nbr_atom.coords)
protein_donors_nbr.append(nbr)
if atom.type in [
'N3+', 'N2+', 'Ng+'
]: # types from mol2 and Chimera, ref: http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/idatm.html
protein_salt_plus.append(atom.coords)
protein_salt_plus_res.append(
residue.name[:3] + str(residue.OBResidue.GetNum()))
if atom.OBAtom.IsHbondAcceptor():
protein_acceptors.append(atom.coords)
protein_acceptors_res.append(
residue.name[:3] + str(residue.OBResidue.GetNum()))
nbr = []
