def pydock(wrkdir):
"""
Calculate pyDock scoring.
[1] T. M.-K. Cheng, T. L. Blundell, and J. Fernandez-Recio,
"pyDock: Electrostatics and desolvation for effective scoring of
rigid-body protein–protein docking", Proteins: Structure,
Function, and Bioinformatics, vol. 68, no. 2, pp. 503-515, 2007.
"""
time_start = timer()
log.info("Getting pyDock scoring...")
basepath = os.getcwd()
os.chdir(wrkdir)
rec = ppdx.Pdb('receptor.pdb')
rec.set_chain('A')
rec.write('receptorA.pdb')
lig = ppdx.Pdb('ligand.pdb')
lig.set_chain('B')
lig.write('ligandB.pdb')
xavg = np.mean([atom.x for atom in lig.atoms])
yavg = np.mean([atom.y for atom in lig.atoms])
zavg = np.mean([atom.z for atom in lig.atoms])
with open("pydock.ini", 'w') as fp:
fp.write("[receptor]\n")
fp.write("pdb = receptorA.pdb\n")
fp.write("mol = %s\n" % ('A'))
fp.write("newmol = %s\n" % ('A'))
fp.write("\n")
fp.write("[ligand]\n")
fp.write("pdb = ligandB.pdb\n")
fp.write("mol = %s\n" % ('B'))
fp.write("newmol = %s\n" % ('B'))
ret = ppdx.tools.execute("%s pydock setup >pydock_setup.out 2>&1" %
(os.path.join(ppdx.PYDOCK, 'pyDock3')))
if ret != 0:
os.chdir(basepath)
raise ValueError("pyDock setup failed!")
with open("pydock.rot", 'w') as fp:
fp.write("1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 %f %f %f 1\n" %
(xavg, yavg, zavg))
ret = ppdx.tools.execute("%s pydock dockser >pydock_dockser.out 2>&1" %
(os.path.join(ppdx.PYDOCK, 'pyDock3')))
if ret != 0:
os.chdir(basepath)
raise ValueError("pyDock dockser failed!")
with open('pydock.ene', 'r') as fp:
line = fp.readlines()[-1]
_, elec, desolv, vdw, total, _ = line.split()
os.chdir(basepath)
desc = {
'pyDock': float(total),
'pyDock_elec': float(elec),
'pyDock_desolv': float(desolv),
'pyDock_vdw': float(vdw)
}
time_end = timer()
desc['>TIME_pyDock'] = time_end - time_start
return desc
def charmify(fname, nsteps=100):
basepath = os.getcwd()
wrkdir, name = os.path.split(fname)
basename = ''.join(name.split('.')[0:-1]) + '-chm'
if os.path.isfile(os.path.join(wrkdir, basename+'.psf')) and os.path.isfile(os.path.join(wrkdir, basename+'.pdb')):
return
else:
log.info("Charmify-ing pdb %s" % (fname))
os.chdir(wrkdir)
pdb = ppdx.Pdb(name)
pdb.fix4charmm()
pdb.chain2segid()
pdb.set_occupancy(1.0)
pdb.set_beta(1.0)
pdb.remove_hydrogens()
chains = pdb.split_by_chain()
nchains = len(chains)
cmd = ppdx.CHARMM
cmd += ' nc=%d ' % nchains
i=1
for ch, pdb in chains.items():
pdb.write("chain_%s.pdb" % (ch.lower()))
cmd += 'c%d=%s ' % (i, ch)
i += 1
cmd += 'name=chain_ out=%s ' % (basename)
cmd += 'nsteps=%d ' % (nsteps)
cmd += 'ffpath=%s ' % (ppdx.FFPATH)
cmd += '-i buildgen.inp >%s 2>&1' % (basename+'.out')
ppdx.link_data('buildgen.inp')
ppdx.link_data('disu.str')
ret = ppdx.tools.execute(cmd)
os.chdir(basepath)
if ret!=0:
raise ValueError("Charmm failed while running < %s > in %s" % (cmd, wrkdir))
def split_complex(wrkdir, nchains):
"""
Split the model-chm.pdb/cor/psf in wrkdir in a ligand-chm.pdb/cor/psf,
receptor-chm.pdb/cor/psf, complex-chm.pdb/cor/psf.
nchains is a tuple containing the number of chains in the receptor
and the number of chains in the ligand.
Assume the chains are called alphabetically (as done by Modeller).
"""
basepath = os.getcwd()
os.chdir(wrkdir)
if not os.path.isfile('ligand-chm.psf') or not os.path.isfile('receptor-chm.psf') or not os.path.isfile('complex-chm.psf') \
or not os.path.isfile('ligand.pdb') or not os.path.isfile('receptor.pdb') or not os.path.isfile('complex.pdb') \
or not os.path.isfile('ligandB.pdb') or not os.path.isfile('receptorA.pdb') or not os.path.isfile('complexAB.pdb'):
log.info('Splitting model-chm.psf in ligand, receptor and complex.')
cpx = ppdx.Pdb(os.path.join(wrkdir, 'model-chm.pdb'))
cpx.segid2chain()
nchains_tot = len(cpx.split_by_chain())
lrec = nchains[0]
llig = nchains[1]
if llig+lrec!=nchains_tot:
raise ValueError('PDB %s contains %d chains, but ligand (%d) and receptor (%d) contain %d' % (wrkdir, nchains_tot, llig, lrec, llig+lrec))
alphabet = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ'
crec = alphabet[:lrec]
clig = alphabet[lrec:nchains_tot]
sele_rec = ' .or. '.join(['segid %s' % (c) for c in crec])
sele_lig = ' .or. '.join(['segid %s' % (c) for c in clig])
ppdx.link_data('extract.inp')
cmd = '%s basename=%s sel="%s" outname=%s ffpath=%s -i extract.inp >%s 2>&1' % (ppdx.CHARMM, 'model-chm', sele_rec, 'receptor-chm', ppdx.FFPATH, 'receptor-chm.out')
ret = ppdx.tools.execute(cmd)
if ret!=0:
raise ValueError("Charmm failed.")
cmd = '%s basename=%s sel="%s" outname=%s ffpath=%s -i extract.inp >%s 2>&1' % (ppdx.CHARMM, 'model-chm', sele_lig, 'ligand-chm', ppdx.FFPATH, 'ligand-chm.out')
ret = ppdx.tools.execute(cmd)
if ret!=0:
raise ValueError("Charmm failed.")
for f in ['cor', 'pdb', 'psf']:
if not os.path.isfile('complex-chm.'+f):
os.symlink('model-chm.'+f, 'complex-chm.'+f)
rec = ppdx.Pdb('receptor-chm.pdb')
lig = ppdx.Pdb('ligand-chm.pdb')
rec.make_standard()
lig.make_standard()
cpx = rec+lig
rec.write('receptor.pdb')
lig.write('ligand.pdb')
cpx.write('complex.pdb')
rec.set_chain('A')
rec.set_segid('A')
lig.set_chain('B')
lig.set_segid('B')
cpx = rec+lig
rec.write('receptorA.pdb')
lig.write('ligandB.pdb')
cpx.write('complexAB.pdb')
os.chdir(basepath)
def rosetta1(target_sequence, template, wrkdir):
"""
Build one model for the complex. Save it in the wrkdir folder.
"""
# Make working directory and enter it
if os.path.isdir(wrkdir):
if os.path.isfile(os.path.join(wrkdir, 'model.pdb')):
return
else:
os.makedirs(wrkdir)
time_start = timer()
log.info("Creating model %s/model.pdb..." % (wrkdir))
basepath = os.getcwd()
os.chdir(wrkdir)
# Create fasta file with target sequence
target_sequence_file = 'target_sequence.fasta'
with open(target_sequence_file, 'w') as fp:
fp.write(">target\n%s\n" % (target_sequence))
# Prepare template
if not os.path.isfile('PDBID.pdb'):
if isinstance(template, ppdx.Pdb):
template.write('PDBID.pdb')
else:
template_path = os.path.join(basepath, template)
if os.path.isfile(template_path):
os.symlink(os.path.join(basepath, template), 'PDBID.pdb')
else:
raise ValueError('Impossible to find template file %s' %
(template))
# Remove / because Bio do not recognize them and rosetta thread neither...
target_sequence = target_sequence.replace('/', '')
template_sequence = ppdx.Pdb("PDBID.pdb").get_sequence().replace('/', '')
# Make alignment
alignment = Bio.pairwise2.align.globalds(target_sequence,
template_sequence,
Bio.SubsMat.MatrixInfo.blosum62,
-10, -0.5)
s1 = alignment[0][0]
s2 = alignment[0][1]
# Write sequence alignemnt in grishnan (?) format
with open('alignment.gri', 'w') as fp:
fp.write("## target PDBID_thread\n")
fp.write("# \n")
fp.write("scores_from_program: 0\n")
fp.write("0 " + s1 + "\n")
fp.write("0 " + s2 + "\n")
fp.write("--\n")
log.info('Running rosetta thread...')
rthread = ppdx.ROSETTABIN+"/partial_thread.static.linuxgccrelease " + \
" -database " + ppdx.ROSETTA + "/main/database" + \
" -in:file:fasta " + target_sequence_file + \
" -in:file:alignment alignment.gri" + \
" -in:file:template_pdb PDBID.pdb" + \
" -ignore_unrecognized_res " + \
" >rosetta_thread.out 2>&1 "
ppdx.tools.execute(rthread)
if not os.path.isfile('PDBID_thread.pdb'):
log.error(
'Rosetta thread failed! Look at the output file rosetta_thread.out'
)
log.info("Running Rosetta script...")
ppdx.link_data('hybridize.xml')
rscript = ppdx.ROSETTABIN + "/rosetta_scripts.static.linuxgccrelease " + \
" -database " + ppdx.ROSETTA + "/main/database " + \
" -in:file:fasta " + target_sequence_file + \
" -parser:protocol hybridize.xml " + \
" -default_max_cycles 200 " + \
" -dualspace " + \
" -restore_talaris_behavior " + \
" -score:set_weights pro_close 0 " + \
" >rosetta_script.out 2>&1 "
ppdx.tools.execute(rscript)
if not os.path.isfile('S_0001.pdb'):
log.error(
'Rosetta hybridize failed! Look at the output file rosetta_script.out'
)
# Done!
if os.path.lexists('model.pdb'):
os.remove('model.pdb')
os.symlink("S_0001.pdb", "model.pdb")
os.chdir(basepath)
log.info("Model %s/model.pdb created!" % (wrkdir))
time_end = timer()
return time_end - time_start
def enm_entropy(fname, K=50.0, cutoff=7.0, spring='constant'):
"""
Given a pdb file, make the Elastic Network Model (ENM),
build its hessian matrix, diagonalize it, and return its entropy.
"""
pdb_all = ppdx.Pdb(fname)
pdb_all.remove_hydrogens()
pdb = [
[atom.x, atom.y, atom.z] for atom in pdb_all.atoms
if (atom.name == 'CB' or (atom.name == 'CA' and atom.resname == 'GLY'))
]
def spring_constant(k, d):
return k / (d * d)
def spring_exponential(k, d, d0=7.0):
return k * exp(-(d / d0)**2)
def spring_overR6(k, d):
return k / (d**4)
if spring == 'constant':
spr = spring_constant
elif spring == 'exponential':
spr = spring_exponential
elif spring == 'overR6':
spr = spring_overR6
else:
log.error(
"Unkown spring method <%s>. Available are: constant, exponential, overR6"
% spring)
return float('nan')
enm = []
natoms = len(pdb)
log.info("Making ENM on %d residues" % (natoms))
for i in range(natoms):
for j in range(i + 1, natoms):
dx = pdb[i][0] - pdb[j][0]
dy = pdb[i][1] - pdb[j][1]
dz = pdb[i][2] - pdb[j][2]
d = sqrt(dx * dx + dy * dy + dz * dz)
if (d < cutoff):
enm.append([i, j, dx, dy, dz, d])
#print("Among all N*(N-1)/2 = %d pairs, %d were selected." % (natoms*(natoms-1)/2, len(enm)))
#print("Building Hessian Matrix")
nfree = natoms * 3
hess = np.zeros((nfree, nfree))
for bond in enm:
i = bond[0]
j = bond[1]
dx = bond[2]
dy = bond[3]
dz = bond[4]
d = bond[5]
tmp = spr(K, d)
for m, v1 in enumerate([dx, dy, dz]):
for n, v2 in enumerate([dx, dy, dz]):
hess[3 * i + m][3 * i + n] += tmp * v1 * v2
hess[3 * j + m][3 * j + n] += tmp * v1 * v2
hess[3 * i + m][3 * j + n] -= tmp * v1 * v2
hess[3 * j + m][3 * i + n] -= tmp * v1 * v2
del enm
#print("Diagonalizing it...")
#eival, eivec = np.linalg.eig(hess)
eival = np.linalg.eigvalsh(hess)
#print("Calculating the entropy")
S = 0
skipped = 0
toout = ""
for i in eival:
if i > 1E-6:
S += 1 + np.log(sqrt(i))
else:
skipped += 1
toout += "%d %f\n" % (skipped, i)
if skipped != 6:
log.warning(
"Warning! There were %d skipped frequencees below 1E-4! Exactly 6 were expected!"
% skipped)
log.warning(toout)
return S
def firedock(wrkdir):
"""
Calculate FireDock scoring.
"""
time_start = timer()
log.info("Getting FireDock scoring...")
basepath = os.getcwd()
os.chdir(wrkdir)
# Remove hydrogens
pdb = ppdx.Pdb('ligand.pdb')
pdb.remove_hydrogens()
pdb.write('ligand_firedock_noh.pdb')
pdb = ppdx.Pdb('receptor.pdb')
pdb.remove_hydrogens()
pdb.write('receptor_firedock_noh.pdb')
# Reduce
fd_reduce = "%s/PDBPreliminaries/reduce.2.21.030604 -DB %s/PDBPreliminaries/reduce_het_dict.txt -OH -HIS -NOADJust -NOROTMET" % (
ppdx.FIREDOCK, ppdx.FIREDOCK)
ret = ppdx.tools.execute(
fd_reduce +
' receptor_firedock_noh.pdb >firedock_receptor.pdb 2>firedock_reduce_receptor.out'
)
if ret != 0:
os.chdir(basepath)
raise ValueError("FireDock reduce receptor failed!")
ret = ppdx.tools.execute(
fd_reduce +
' ligand_firedock_noh.pdb >firedock_ligand.pdb 2>firedock_reduce_ligand.out'
)
if ret != 0:
os.chdir(basepath)
raise ValueError("FireDock reduce ligand failed!")
# Score
with open('firedock.trans', 'w') as fp:
fp.write("1 0.0 0.0 0.0 0.0 0.0 0.0")
ret = ppdx.tools.execute(
"%s/buildFireDockParams.pl firedock_receptor.pdb firedock_ligand.pdb U U Default firedock.trans firedock_build.out 1 50 0.85 0 firedock_parameters.dat >firedock_build.err 2>&1"
% ppdx.FIREDOCK)
if ret != 0:
os.chdir(basepath)
raise ValueError("FireDock build param failed!")
ret = ppdx.tools.execute(
"%s/runFireDock.pl firedock_parameters.dat >firedock.log 2>&1" %
(ppdx.FIREDOCK))
if ret != 0:
os.chdir(basepath)
raise ValueError("FireDock scoring failed!")
# Get values
with open('firedock_build.out.ref', 'r') as fp:
data = fp.readlines()[-1].split('|')
if data[5].strip() == 'glob':
log.warning('Problem with FireDock in %s' % (wrkdir))
data = [float('nan')] * 20
desc = dict()
desc['FireDock'] = float(data[5])
desc['FireDock_aVdW'] = float(data[6])
desc['FireDock_rVdW'] = float(data[7])
desc['FireDock_ACE'] = float(data[8])
desc['FireDock_inside'] = float(data[9])
desc['FireDock_aElec'] = float(data[10])
desc['FireDock_rElec'] = float(data[11])
desc['FireDock_laElec'] = float(data[12])
desc['FireDock_lrElec'] = float(data[13])
desc['FireDock_hb'] = float(data[14])
desc['FireDock_piS'] = float(data[15])
desc['FireDock_catpiS'] = float(data[16])
desc['FireDock_aliph'] = float(data[17])
os.chdir(basepath)
time_end = timer()
desc['>TIME_FireDock'] = time_end - time_start
return desc