def sa(intra=False,rainbow=True):
"""
Superimpose all open models onto the first one.
This may not work well with selections.
Option intra can be set to True to enable intra_fit first, for working with multi-state (nmr) pdbs.
[Thanks to Kyle Beauchamp for this one]
"""
AllObj=cmd.get_names("all")
for x in AllObj:
print(AllObj[0],x)
if intra==True:
cmd.intra_fit(x)
if rainbow==True:
cmd.util.chainbow(x)
cmd.align(x,AllObj[0])
cmd.zoom()
def sa(intra=False, rainbow=True):
"""
Superimpose all open models onto the first one.
This may not work well with selections.
Option intra can be set to True to enable intra_fit first, for working with multi-state (nmr) pdbs.
[Thanks to Kyle Beauchamp for this one]
"""
AllObj = cmd.get_names("all")
for x in AllObj:
print(AllObj[0], x)
if intra == True:
cmd.intra_fit(x)
if rainbow == True:
cmd.util.chainbow(x)
cmd.align(x, AllObj[0])
cmd.zoom()
def testIntraFit(self):
cmd.fragment("gly", "m1")
cmd.create("m1", "m1", 1, 2)
rms_list = cmd.intra_fit("m1")
self.assertArrayEqual(rms_list, [-1.0, 0.0])
rms_list = cmd.intra_rms("m1")
self.assertArrayEqual(rms_list, [-1.0, 0.0])
rms_list = cmd.intra_rms_cur("m1")
self.assertArrayEqual(rms_list, [-1.0, 0.0])
def testIntraFit(self):
cmd.fragment("gly", "m1")
cmd.create("m1", "m1", 1, 2)
rms_list = cmd.intra_fit("m1")
self.assertArrayEqual(rms_list, [-1.0, 0.0])
rms_list = cmd.intra_rms("m1")
self.assertArrayEqual(rms_list, [-1.0, 0.0])
rms_list = cmd.intra_rms_cur("m1")
self.assertArrayEqual(rms_list, [-1.0, 0.0])
def clean(pose):
"""Deletes everything and load the validated models"""
cmd.delete('all')
cmd.load('%s_Ca.pdb' % pose)
cmd.load('%s_Cb.pdb' % pose)
cmd.load('%s_CaCb.pdb' % pose)
cmd.intra_fit('%s_Ca' % pose, 0, quiet=1)
cmd.intra_fit('%s_Cb' % pose, 0, quiet=1)
cmd.intra_fit('%s_CaCb' % pose, 0, quiet=1)
cmd.dss('all')
cmd.disable('%s_Ca' % pose)
cmd.disable('%s_Cb' % pose)
def getRMSD(directory="temp"):
# If the directory exists
if os.path.exists(directory):
files = [
f for f in listdir(directory) if isfile(join(directory, f)) and (
f.split(".")[-1] == "log" or f.split(".")[-1] == "out")
and len(f.split(".")) > 1
]
# load all the pdb files
for file in files:
filesplit = file.split(".")
full_file_name = join(directory, "%s_OPT.pdb" % filesplit[0])
cmd.load(full_file_name, "%s_copy" % cmd.get_object_list()[0])
# calculate for the reference object
RMSD = cmd.intra_fit("%s_copy" % cmd.get_object_list()[0], 0)
#print RMSD
energies = []
for i in range(len(files)):
out = getoutData(join(directory, files[i]))
energies.append(out.ENERGY)
# scale and convert hartree to J/mol. Then calculate Boltzman distribution
hartree_to_Jmol = 2600 * 1000 # hartree to J/mol
Jmol_to_kcalmol = 0.24 / 1000 # J/mol to kcal/mol
min_energy = min(energies)
for i in range(len(energies)):
energies[i] = (energies[i] -
min_energy) * hartree_to_Jmol * Jmol_to_kcalmol
#write csv
output = open("RMSD.csv", 'w')
output.write("Structure;Energy;RMSD\n")
message = "%s;%.4f;%.4f\n" % (
files[0], energies[0], 0
) # RMSD = -1 for the same structure by default, but in our case 0 is more meningful
output.write(message)
for i in range(1, len(files)):
message = "%s;%.4f;%.4f\n" % (files[i], energies[i], RMSD[i])
output.write(message)
# terminate
output.close()
else:
print("FATAL ERROR: Specified directory doesn't exist [%s]" %
directory)
def filter_opt(directory="temp", setup_path="default", verbose="False"):
# If the directory exists
if os.path.exists(directory):
# Log generation
log = Log()
log.write(
"\n\n########################################\n########### F I L T E R - O ############\n########################################\n\n"
)
#verbose
if verbose.lower() in ['true', '1', 't', 'y', 'yes']: verbose = True
else: verbose = False
# Retrieve all the files in the directory which are output with .log extension from gaussian or .out extension from mopac
files = [
f for f in listdir(directory) if isfile(join(directory, f)) and (
f.split(".")[-1] == "log" or f.split(".")[-1] == "out")
and len(f.split(".")) > 1
]
if len(files) > 0:
# get setup.ini
setup = Setup(log, setup_path)
if setup.isLoaded() == True:
# if optimisation, remove identical conformers
if setup.scf != "1SCF":
log.write(
"\n\nRemove identical structure with an RMSD cutoff of: %s Angstroms "
% setup.rmsd_cutoff)
#prepare bins
bins = [
] #first of the bin is the median of the bin, ie the representation of the bin
energies = []
# Retrieve energy and create associated pdb
for i in range(len(files)):
fileData = getoutData(join(directory, files[i]))
#there is an energy, there was no error
if hasattr(fileData, "ENERGY"):
export_pdb(directory, files[i], fileData)
energies.append(fileData.ENERGY) # hartree
else: #error: the optimisation probably failed for this conformer, we remove it and trigger a warning
os.remove(join(directory, files[i]))
log.write(
"WARNING: No energy was found. Removed from conformers [%s]"
% join(directory, files[i]))
#reload
files = [
f for f in listdir(directory)
if isfile(join(directory, f)) and
(f.split(".")[-1] == "log" or f.split(".")[-1] == "out"
) and len(f.split(".")) > 1
]
# load all the pdb files
for file in files:
filesplit = file.split(".")
full_file_name = join(directory,
"%s_OPT.pdb" % filesplit[0])
cmd.load(full_file_name,
"%s_copy" % cmd.get_object_list()[0])
# calculate for each file
for i in range(len(files)):
log.write(
"\n%s_copy__%s RMSD:" %
(cmd.get_object_list()[0], i + 1), verbose)
RMSD = cmd.intra_fit(
"%s_copy" % cmd.get_object_list()[0], i + 1)
log.write("%s" % RMSD, verbose)
min_id = -1
min_rmsd = setup.rmsd_cutoff
for j in range(len(bins)):
current_rmsd = RMSD[bins[j][0]]
# if under a certain cut-off, considers the same structure
if current_rmsd < setup.rmsd_cutoff:
# put the structure in the bin with the minimum cutoff
if current_rmsd < min_rmsd:
min_id = j
min_rmsd = current_rmsd
# check results, and do accordingly
if min_id == -1:
bins.append([
i
]) # doesn't belong to a bin, create a new bin
log.write(
"Create a new bin (#%s) for [%s]" %
((len(bins) - 1), files[i]), verbose)
else:
bins[min_id].append(i) #belongs to bin[min_id]
log.write(
"[%s] Added to bin #%s with RMSD of %-6.4f Angstroms"
% (files[i], min_id, min_rmsd), verbose)
#finally, filter by keeping one structure for each bin
for i in range(len(bins)):
min_energy = bins[i][0]
for j in range(1, len(bins[i])):
# keep the smallest energy conformation
if energies[bins[i][j]] < energies[min_energy]:
try: # remove previous files
filename = files[min_energy].split(".")[0]
# remove *.pdb, *.out or *.log
os.remove(
join(directory,
"%s_OPT.pdb" % filename))
os.remove(
join(directory, files[min_energy]))
if setup.software == "MOPAC": # remove *.mop, *.arc
os.remove(
join(directory,
"%s.mop" % filename))
os.remove(
join(directory,
"%s.arc" % filename))
elif setup.software == "GAUSSIAN":
os.remove(
join(directory,
"%s.com" % filename))
except:
log.write(
"WARNING: Failed to remove [%s]" %
join(directory, files[min_energy]))
pass
min_energy = bins[i][j]
else:
try: # remove these files
filename = files[bins[i][j]].split(".")[0]
# remove *.pdb, *.out or *.log
os.remove(
join(directory,
"%s_OPT.pdb" % filename))
os.remove(
join(directory, files[bins[i][j]]))
if setup.software == "MOPAC": # remove *.mop, *.arc
os.remove(
join(directory,
"%s.mop" % filename))
os.remove(
join(directory,
"%s.arc" % filename))
elif setup.software == "GAUSSIAN":
os.remove(
join(directory,
"%s.com" % filename))
except:
log.write(
"WARNING: Failed to remove [%s]" %
join(directory, files[bins[i][j]]))
pass
#remove loaded structures from Pymol
cmd.delete("%s_copy" % cmd.get_object_list()[0])
# create output
output = open("energies.txt", 'w')
output.write(
"*******************************************************************************\n"
)
output.write(
"** **\n"
)
output.write(
"** E N E R G I E S **\n"
)
output.write(
"** **\n"
)
output.write(
"*******************************************************************************\n"
)
output.write(
"* CALCULATION WITH %-30s *\n"
% setup.software)
if setup.software == "MOPAC":
output.write("* MOPAC EXECUTIVE PATH: %-53s *\n" %
setup.exe)
output.write("* using force-field: %-56s *\n" % setup.SE)
keywords = "%s charge=%s %s" % (setup.SE, setup.charge,
setup.scf)
output.write("* %-75s *\n" % keywords)
elif setup.software == "GAUSSIAN":
output.write(
"* %%mem=%2.fGB *\n"
% setup.memories)
output.write(
"* %%nprocshared=%2.f *\n"
% setup.procsshared)
output.write(
"* # opt=(maxcycles=160) freq=noraman %-40s *\n" %
setup.leveloftheory)
output.write(
"* %2.f %2.f *\n"
% (setup.charge, setup.spin))
if setup.singlepointcalculation != "":
output.write(
"* --Link1-- *\n"
)
output.write("* # geom=check guess=read %-51s *\n" %
setup.singlepointcalculation)
output.write(
"* %2.f %2.f *\n"
% (setup.charge, setup.spin))
output.write(
"* Clusterisation cut-off was set at: %4.2f Angstroms *\n"
% setup.rmsd_cutoff)
output.write(
"* Angle count: %-3.f | Atomic model: %-5s | Temperature: %4.f K *\n"
% (setup.angle_count, setup.radii, setup.Temperature))
output.write(
"*******************************************************************************\n"
)
output.write(" File created on %s\n" %
str(datetime.date.today()))
output.write(
"\n Structure Etot (Hartree)")
files = [
f for f in listdir(directory)
if isfile(join(directory, f)) and (
f.split(".")[-1] == "log" or f.split(".")[-1] == "out")
and len(f.split(".")) > 1
]
for filename in files:
out = getoutData(join(directory, filename))
message = "\n %-30s" % filename + " %12.6f" % out.ENERGY
output.write(message)
# terminate
output.close()
log.write(
"----------------------------\n---- Normal Termination ----\n----------------------------\n"
)
log.finalize()
return True
else:
log.write(
"Error: Failed to load setup.ini in [%s]. Fix it to continue."
% setup_path)
return False
else:
log.write("Error: No file to use in the directory [%s]" %
directory)
return False
else:
print("FATAL ERROR: Specified directory doesn't exist [%s]" %
directory)
return False
model = cmd.get_model('system')
#for atom in model.atom:
# print "%8d %4s %3s %5d %8.3f %8.3f %8.3f" % (atom.index, atom.name, atom.resn, int(atom.resi), atom.coord[0], atom.coord[1], atom.coord[2])
#pymol.finish_launching()
cmd.viewport(width,height)
#niterations = 10 # DEBUG
# Load trajectory
cmd.load_traj(trajectory_temporary_filename, object='system')
# Align all states
if solute == 'dna':
cmd.intra_fit('name P') # DNA
else:
cmd.intra_fit('solute') # protein
# Zoom viewport
cmd.zoom('solute')
cmd.orient('solute')
# Create one-to-one mapping between states and frames.
cmd.mset("1 -%d" % nframes)
# Delete first frame
cmd.mdelete("1")
# Render movie
def mcm(pose, mc_steps, SASA, randomize):
################################# MCM Parameters ##########################
T = 300. # Temperature
k = 0.0019872041 # Boltzmann constant
kT = k * T
# probability to sample phi, psi or chi
angles_prob = [1 / 3, 1 / 3, 1 / 3]
accepted = 0
##########################################################################
#
first, last = pose_from_pdb(pose)
if first or last:
print('Starting MCM')
from energy import minimize, set_sasa, get_sasa
sus_updates = cmd.get('suspend_updates')
cmd.set('suspend_updates', 'on')
# uncomment for debugging
cmd.feedback('disable', 'executive', 'everything')
pdb_conect = cmd.get('pdb_conect_all')
cmd.set('pdb_conect_all', 1)
glyco_bonds = get_glyco_bonds(first, last)
con_matrix = writer(glyco_bonds)
# Remove previous pdb files
prev_files = glob.glob('mcm_*.pdb')
for prev_file in prev_files:
os.remove(prev_file)
# set all paramenters for sasa-energy computation
if SASA:
params, points, const = set_sasa(n=1000)
# randomize initial conformation
if randomize:
for i in range(len(con_matrix) - 1):
bond = con_matrix[i]
angle_values = np.random.uniform(-180, 180, size=2)
set_psi(pose, bond, angle_values[0])
set_phi(pose, bond, angle_values[1])
for i in range(6):
set_chi(pose, bond)
# minimize energy of starting conformation and save it
NRG_old = minimize(pose, nsteps=5000, rigid_geometry=False)
NRG_min = NRG_old
cmd.save('mcm_%08d.pdb' % accepted)
# start MCM routine
fd = open("mcm_log.txt", "w")
print('# iterations remaining = %s' % (mc_steps))
for i in range(1, mc_steps + 1):
if i % (mc_steps // 10) == 0:
print('#remaining iterations = %s' % (mc_steps - i))
if True:
sample_uniform(pose, con_matrix, angles_prob)
NRG_new = minimize('tmp', nsteps=100, rigid_geometry=False)
if SASA:
solvatation_nrg = get_sasa(params,
points,
const,
selection='all',
probe=0)[0]
NRG_new = NRG_new + solvatation_nrg
if NRG_new < NRG_old:
NRG_old = NRG_new
fd.write('%8d%10.2f\n' % (accepted, NRG_new))
cmd.copy(pose, 'tmp')
cmd.delete('tmp')
cmd.save('mcm_%08d.pdb' % accepted)
accepted += 1
else:
delta = np.exp(-(NRG_new - NRG_old) / (kT))
if delta > np.random.uniform(0, 1):
NRG_old = NRG_new
fd.write('%8d%10.2f\n' % (accepted, NRG_new))
cmd.copy(pose, 'tmp')
cmd.delete('tmp')
cmd.save('mcm_%08d.pdb' % accepted)
accepted += 1
cmd.delete('tmp')
if NRG_new < NRG_min:
NRG_min = NRG_new
cmd.save('mcm_min.pdb')
fd.close()
cmd.delete('all')
print('Savings all accepted conformations on a single file')
de_builds = cmd.get('defer_builds_mode')
cmd.set('defer_builds_mode', 5)
for i in range(0, accepted):
cmd.load('mcm_%08d.pdb' % i, 'mcm_trace')
cmd.save('mcm_trace.pdb', 'all', state=0)
cmd.delete('all')
cmd.load('mcm_trace.pdb')
cmd.intra_fit('mcm_trace')
print('MCM completed')
# restore settings
cmd.set('suspend_updates', sus_updates)
cmd.set('pdb_conect_all', pdb_conect)
cmd.set('defer_builds_mode', de_builds)
# asdf
for a in xrange(1, 14):
print a,
print "%8.3f" % cmd.fit("ref and resi %d" % a, "trg"),
print "%8.3f" % cmd.rms("ref", "trg"),
print "%8.3f" % cmd.rms_cur("ref", "trg")
cmd.frame(10)
print "%8.3f" % cmd.fit("ref", "trg")
for a in xrange(1, 14):
print a,
print "%8.3f" % cmd.fit("ref and resi %d" % a, "trg"),
print "%8.3f" % cmd.rms("ref", "trg"),
print "%8.3f" % cmd.rms_cur("ref", "trg")
a = 1
print "%8.3f" % cmd.fit("ref", "trg")
for b in xrange(1, 11):
cmd._dump_floats(cmd.intra_fit("trg and resi %d" % a, b))
cmd._dump_floats(cmd.intra_rms("trg", b))
cmd._dump_floats(cmd.intra_rms_cur("trg", b))
cmd.do("intra_fit (trg),10")
cmd.do("intra_fit (trg and resi 1),10")
cmd.do("intra_rms (trg),10")
cmd.do("intra_rms_cur (trg),10")
print "END-LOG"
for a in xrange(1, 14):
print a,
print "%8.3f" % cmd.fit("ref and resi %d" % a, "trg"),
print "%8.3f" % cmd.rms("ref", "trg"),
print "%8.3f" % cmd.rms_cur("ref", "trg")
cmd.frame(10)
print "%8.3f" % cmd.fit("ref", "trg")
for a in xrange(1, 14):
print a,
print "%8.3f" % cmd.fit("ref and resi %d" % a, "trg"),
print "%8.3f" % cmd.rms("ref", "trg"),
print "%8.3f" % cmd.rms_cur("ref", "trg")
a = 1
print "%8.3f" % cmd.fit("ref", "trg")
for b in xrange(1, 11):
cmd._dump_floats(cmd.intra_fit("trg and resi %d" % a, b))
cmd._dump_floats(cmd.intra_rms("trg", b))
cmd._dump_floats(cmd.intra_rms_cur("trg", b))
cmd.do("intra_fit (trg),10")
cmd.do("intra_fit (trg and resi 1),10")
cmd.do("intra_rms (trg),10")
cmd.do("intra_rms_cur (trg),10")
print "END-LOG"
else:
stop = -1
args.frames = range(1, nframes + 1)
for chunkid, chunk in enumerate(chunks):
start, stop = min(chunk), max(chunk)
cmd.reinitialize()
cmd.load(args.top, 'inp')
cmd.load_traj(args.traj,
'inp',
state=1,
start=start,
stop=stop,
selection=selection)
if args.align is not None:
if len(chunks) == 1:
rmsds = cmd.intra_fit(selection, args.align)
rmsds[rmsds == -1.] = 0.
outrmsdfile = f"{os.path.splitext(args.out)[0]}_rmsd.txt"
numpy.savetxt(outrmsdfile, rmsds, fmt="%.4f")
states = numpy.where(numpy.isin(chunk, args.frames))[0] + 1
nstates = cmd.count_states('inp')
if len(states) == nstates:
cmd.create('out', selection=selection, source_state=0, target_state=-1)
else:
for s in states:
sys.stdout.write(f'Getting state {s}/{max(states)}\r')
sys.stdout.flush()
cmd.create('out',
selection=selection,
source_state=s,
target_state=-1)
def mcm(pose, mc_steps, SASA, randomize):
################################# MCM Parameters ##########################
T = 300. # Temperature
k = 0.0019872041 # Boltzmann constant
angles_prob = [1/3, 1/3, 1/3] # probability to sample phi, psi or chi
accepted = 0
############################################################################
#
first, last = pose_from_pdb(pose)
if first or last:
print('Starting MCM')
from energy import minimize, set_sasa, get_sasa
cmd.set('suspend_updates', 'on')
cmd.feedback('disable', 'executive', 'everything') ##uncomment for debugging
cmd.set('pdb_conect_all', 1)
glyco_bonds = get_glyco_bonds(first, last)
con_matrix = writer(glyco_bonds)
# Remove previous pdb files
prev_files = glob.glob('mcm_*.pdb')
for prev_file in prev_files:
os.remove(prev_file)
# set all paramenters for sasa-energy computation
if SASA:
params, points, const = set_sasa(n=1000)
## randomize initial conformation
if randomize:
for i in range(len(con_matrix)-1):
bond = con_matrix[i]
angle_values = np.random.uniform(-180, 180, size=2)
set_psi(pose, bond, angle_values[0])
set_phi(pose, bond, angle_values[1])
for i in range(6):
set_chi(pose, bond)
# minimize energy of starting conformation and save it
NRG_old = minimize(pose, nsteps=5000, rigid_geometry=False)
NRG_min = NRG_old
cmd.save('mcm_%08d.pdb' % accepted)
## start MCM routine
fd = open("mcm_log.txt", "w")
print('# iterations remaining = %s' % (mc_steps))
for i in range(1, mc_steps+1):
if i % (mc_steps//10) == 0:
print('#remaining iterations = %s' % (mc_steps-i))
if True:
sample_uniform(pose, con_matrix, angles_prob)
NRG_new = minimize('tmp', nsteps=100, rigid_geometry=False)
if SASA:
solvatation_nrg = get_sasa(params, points, const, selection='all',
probe=0)[0]
NRG_new = NRG_new + solvatation_nrg
if NRG_new < NRG_old:
NRG_old = NRG_new
fd.write('%8d%10.2f\n' % (accepted, NRG_new))
cmd.copy(pose, 'tmp')
cmd.delete('tmp')
cmd.save('mcm_%08d.pdb' % accepted)
accepted += 1
else:
delta = np.exp(-(NRG_new-NRG_old)/(T*k))
if delta > np.random.uniform(0, 1):
NRG_old = NRG_new
fd.write('%8d%10.2f\n' % (accepted, NRG_new))
cmd.copy(pose, 'tmp')
cmd.delete('tmp')
cmd.save('mcm_%08d.pdb' % accepted)
accepted += 1
cmd.delete('tmp')
if NRG_new < NRG_min:
NRG_min = NRG_new
cmd.save('mcm_min.pdb')
fd.close()
cmd.delete('all')
print('Savings all accepted conformations on a single file')
cmd.set('defer_builds_mode', 5)
for i in range(0, accepted):
cmd.load('mcm_%08d.pdb' % i, 'mcm_trace')
cmd.save('mcm_trace.pdb', 'all', state=0)
cmd.delete('all')
cmd.load('mcm_trace.pdb')
cmd.intra_fit('mcm_trace')
print(' MCM completed')
cmd.set('suspend_updates', 'off')
def avg_states(object='all',
object_sel=None,
first=1,
last=0,
newobj=None,
fitverdict='no',
verb=0,
pairs=1,
writefiles=1):
"""
ARGUMENTS:
object (string [defaults to 'all']):
Starting PyMOL molecular object consisting of >1 states to be averaged
object_sel (string [defaults to "name CA"]):
An optional subset of atoms on which to operate (e.g., "name CA and
resi 20-50")
first (int [defaults to 1]):
number of the first state to include in averaging
last (int [defaults to last state]):
Number of the last state to include in averaging. A value of '0' (zero)
means use the last state in the object.
newobj (string [defaults to name of object with string "_avg_AtoZ"
appended,
where A and Z are the numbers of the first and last frames included
in the averaging]):
Desired name of the new output pymol object (i.e., averaged structure)
fitverdict (string [defaults to "no", any value != "no" is taken as a
"yes"]):
Use the pymol function intra_fit() to fit all the states of "object &
object_sel"
before calculating the average structure and RMSDs??
'no' = NO, !'no' = yes
verb (int [defaults to 0]):
Verbosity level of output. 0 = quiet, !0 = verbose (e.g., write
position-specific
RMSDs to standard output).
pairs (int [default 1]):
Also calculate average pairwise RMSD for each residue position??
0 = no, !0 = yes
writefiles (int [default 1]):
Write the calculated RMSD data to plaintext files?? 0 = no, !0 = yes
NOTES:
The state specified as 'first' is taken as the reference state for
the (optional)
intra_fit() alignment. If no selection is given, the string "and name
CA" will be
appended to the object and the averaging and rmsd calculations will be
restricted
to "object and name CA" (i.e., C-alpha atoms). To avoid this default
behavior,
specify a valid pymol atom selection for the object_sel argument (e.g.,
"name CA"
or "name CA and resi 20-50" or whatever).
"""
pair_file = None
object = str(object)
object_sel = str(object_sel)
first = int(first)
last = int(last)
num_states_tot = cmd.count_states(object)
if last < 1:
last = num_states_tot
num_states2avg = last - first + 1
if newobj is None or newobj == '':
newobj = "%s_avg_%dto%d" % (object, first, last)
cmd.create(newobj, object, first, 1)
if writefiles:
print('%s' % ('-' * 80))
datfileprefix = '%s_%dto%d' % (replace(object, ' ', '_'), first, last)
avg_rmsd_file = datfileprefix + '.avg_rmsd.dat'
avg_file = open(avg_rmsd_file, 'w')
print('Opened "%s" file for writing residue-specific RMSDs to '
'average structure...' % avg_rmsd_file)
if pairs:
pair_rmsd_file = datfileprefix + '.pair_rmsd.dat'
pair_file = open(pair_rmsd_file, 'w')
print('Opened "%s" file for writing averaged residue-specific '
'pairwise RMSDs...' % pair_rmsd_file)
print('%s' % ('-' * 80))
else:
# just do this instead of evaluating conditionals each time
avg_file = open('/dev/null', 'w')
if pairs:
pair_file = open('/dev/null',
'w') # through the atom loops (below)...
obj_sel_string = (lambda o, sel: (o and sel and "%s and %s" % (o, sel)) or
(o and not sel and "%s and name CA" % o))(object,
object_sel)
newobj_sel_string = (lambda o, sel: (o and sel and "%s and %s" %
(o, sel)) or
(o and not sel and "%s and name CA" % o))(newobj,
object_sel)
print('%s' % ('-' * 80))
print(
'Averaging %d states [%d, %d] of object "%s" (%d total states) '
'to get new single-state object "%s"...' %
(num_states2avg, first, last, obj_sel_string, num_states_tot, newobj))
print('%s' % ('-' * 80))
tmpobject = '%s_tmp4avg_%dto%d' % (object, first, last)
i = 0
for eachstate in range(first, last + 1):
i += 1 # because pymol states are indexed starting from 1 (not 0)
cmd.create(tmpobject, obj_sel_string, eachstate, i)
if fitverdict != 'no':
print("-> proceeding WITH FITTING to reference state...")
tmpobject_intrafit_rmsds = cmd.intra_fit(tmpobject)
if verb:
print(' intrafit_rmsds = ', )
print(tmpobject_intrafit_rmsds)
else:
print("-> proceeding WITHOUT FITTING to reference state...")
# create an atom index map between original (complete) object and subset
# to be averaged:
atindex_map = [None]
for at in cmd.get_model(newobj_sel_string).atom:
atindex_map.append(at.index)
if verb:
print("-> atom index_map = ", )
print(atindex_map)
newobj_chempy = cmd.get_model(tmpobject)
for at in newobj_chempy.atom:
this_at_idx = at.index
sum_x = sum_y = sum_z = 0.0
# avg_x = avg_y = avg_z = 0.0
state_coords = []
# rmsd_sum = rmsd = 0.0
rmsd_sum = 0.0
for s in range(1, num_states2avg + 1):
this_x = cmd.get_model(tmpobject, s).atom[this_at_idx - 1].coord[0]
this_y = cmd.get_model(tmpobject, s).atom[this_at_idx - 1].coord[1]
this_z = cmd.get_model(tmpobject, s).atom[this_at_idx - 1].coord[2]
sum_x += this_x
sum_y += this_y
sum_z += this_z
state_coords.append([this_x, this_y, this_z])
avg_x = sum_x / num_states2avg
avg_y = sum_y / num_states2avg
avg_z = sum_z / num_states2avg
cmd.alter_state(
1, '{0:s} and id {1:d}'.format(newobj, atindex_map[this_at_idx]),
'x=%f' % avg_x)
cmd.alter_state(
1, '{0:s} and id {1:d}'.format(newobj, atindex_map[this_at_idx]),
'y=%f' % avg_y)
cmd.alter_state(
1, '{0:s} and id {1:d}'.format(newobj, atindex_map[this_at_idx]),
'z=%f' % avg_z)
# position-specific RMSDs with respect to average structure:
for somept in state_coords:
rmsd_sum += pow(calcDist(somept, [avg_x, avg_y, avg_z]), 2.0)
rmsd = sqrt(rmsd_sum / num_states2avg)
# DEBUG:
# DEBUG print('newobj = %sel / id = %d / b = %0.3f'
# %(newobj,atindex_map[this_at_idx],rmsd)
cmd.alter(
'{0:s} and id {1:d}'.format(newobj, atindex_map[this_at_idx]),
'b=%0.3f' % rmsd)
if verb:
print('"%s" index = %d' % (obj_sel_string, this_at_idx))
print('rmsd = %0.3f' % rmsd)
avg_file.write('{0:d}\t{1:0.3f}\n'.format(atindex_map[this_at_idx],
rmsd))
# position-specific RMSDs averaged pairwise over the bundle:
if pairs:
for at in newobj_chempy.atom:
this_at_idx = at.index
loop_counter = 0
running_sum = 0.0
for s1 in range(1, num_states2avg + 1):
for s2 in range(s1 + 1, num_states2avg + 1):
# print('computing position %d, %d x %d'%(this_at_idx,s1,s2)
pos1 = [
cmd.get_model(tmpobject,
s1).atom[this_at_idx - 1].coord[0],
cmd.get_model(tmpobject,
s1).atom[this_at_idx - 1].coord[1],
cmd.get_model(tmpobject,
s1).atom[this_at_idx - 1].coord[2]
]
pos2 = [
cmd.get_model(tmpobject,
s2).atom[this_at_idx - 1].coord[0],
cmd.get_model(tmpobject,
s2).atom[this_at_idx - 1].coord[1],
cmd.get_model(tmpobject,
s2).atom[this_at_idx - 1].coord[2]
]
loop_counter += 1
running_sum += pow(calcDist(pos1, pos2), 2.0)
pairwise_rmsd = sqrt(running_sum / loop_counter)
if verb:
print('calculated pairwise RMSD for {0:d} pairs at '
'position {0:d} = {1:0.3f}'.format(
loop_counter, atindex_map[this_at_idx],
pairwise_rmsd))
pair_file.write('{0:d}\t{1:0.3f}\n'.format(
atindex_map[this_at_idx], pairwise_rmsd))
if writefiles:
avg_file.close()
if pairs:
pair_file.close()
cmd.delete(tmpobject)
return newobj
]
import numpy as np
bb_rmsd_mean = np.mean(bb_rmsds)
bb_rmsd_std = np.std(bb_rmsds)
ha_rmsd_mean = np.mean(ha_rmsds)
ha_rmsd_std = np.std(ha_rmsds)
print('\n\n* POKY ENSEMBLE RMSD')
print('* PDB FILE: ' + pdbpath)
print('* Ordered residues by CYRANGE: ' + ordered)
print('MODEL BB RMSD HA RMSD')
for i in range(2, 21):
m = '%2d <-> 1 ' % (i)
m += '%6.3f ' % (bb_rmsds[i - 2])
m += '%6.3f ' % (ha_rmsds[i - 2])
print(m)
print('--------------------------------------------')
print('MEAN RMSD : %6.3f %6.3f' % (bb_rmsd_mean, ha_rmsd_mean))
print('STDEV RMSD: %6.3f %6.3f' % (bb_rmsd_std, ha_rmsd_std))
if s.show_message_yes_no('Save',
'Do you want to save the superimposed structure?'):
out_path = s.save_filedialog('Save as',
'PDB (*.pdb);; mmCIF (*.cif);; Any (*)', '')
if out_path == '':
raise SystemExit
cmd.intra_fit('for_rmsd & i. %s & n. N*+C*+O*+S*' % (oreg))
cmd.save(out_path, 'for_rmsd', state=0)
s.show_message('Finished', 'Finished.')
import argparse
# argparse.ArgumentParser(prog=None, usage=None, description=None, epilog=None, parents=[], formatter_class=argparse.HelpFormatter, prefix_chars='-', fromfile_prefix_chars=None, argument_default=None, conflict_handler='error', add_help=True, allow_abbrev=True, exit_on_error=True)
parser = argparse.ArgumentParser(description='')
# parser.add_argument(name or flags...[, action][, nargs][, const][, default][, type][, choices][, required][, help][, metavar][, dest])
parser.add_argument('-p', '--pdb')
parser.add_argument('-t', '--traj')
parser.add_argument('-s', '--sel', default='all')
parser.add_argument('-r', '--refs', nargs='+')
parser.add_argument('-o', '--out', default='rmsds.txt')
args = parser.parse_args()
cmd.load(args.pdb, 'trajin')
cmd.load_traj(args.traj, 'trajin', state=1)
nframes = cmd.count_states('trajin')
print(f'Number of frames: {nframes}')
for ref in args.refs:
print(f'Loading {ref}')
cmd.load(ref, 'trajin')
cmd.remove('hydro and trajin')
nref = len(args.refs)
rmsds_all = []
for ind, frameref in enumerate([nframes + i for i in range(1, nref + 1)]):
print(f'Computing RMSD for {args.refs[ind]}')
rmsds = cmd.intra_fit(f'trajin and {args.sel}', state=frameref)[:nframes]
rmsds_all.append(rmsds)
rmsds_all = np.asarray(rmsds_all, dtype=object).T
min_rmsd = rmsds_all.min(axis=1)
args_min = rmsds_all.argmin(axis=1)
args_min = np.asarray(args.refs)[args_min]
np.savetxt(args.out, np.c_[rmsds_all, min_rmsd, args_min], fmt=' '.join(['%.4f', ] * (nref + 1)) + ' %s', header=' '.join([f'#{ref}' for ref in args.refs]) + ' #min_rmsd' + ' #min_rmsd_ref', comments='')
cmd.rewind()
cmd.reset()
cmd.delete("all")
# Retain order and IDs for imported atoms
cmd.set("retain_order", 1)
cmd.set("pdb_retain_ids", 1)
# Load PDB file
cmd.load(pdb_filename, "complex")
# Recognize helices and sheets
cmd.dss()
# Align everything
cmd.intra_fit("all")
# Set up display
cmd.remove("resn WAT") # remove waters
cmd.select("receptor", "(not resn MOL) and (not resn WAT) and (not hydrogen)")
cmd.select("ligand", "resn MOL")
cmd.deselect()
cmd.hide("all")
cmd.show("cartoon", "receptor")
cmd.show("spheres", "ligand")
util.cbay("ligand")
cmd.color("green", "receptor")
# Show surface
# cmd.show('surface', 'receptor')
# cmd.set('surface_color', 'white')
cmd.create('complex', 'complex', 1, iteration+1)
cmd.alter_state(iteration+1, 'complex', '(x,y,z) = xyz_iter.next()', space=locals())
#for pdb_index in range(natoms):
#if (pdb_index % 100)==0: print pdb_index
#model_index = model_mapping[pdb_index]
#model.atom[model_index].coord = (10 * ncfile.variables['positions'][iteration, replica, pdb_index, :]).squeeze().tolist()
#for k in range(3):
# model.atom[model_index].coord[k] = float(ncfile.variables['positions'][iteration, replica, pdb_index, k]) * 10.0 # convert to angstroms
#cmd.load_model(model, 'complex', state=iteration+1)
#cmd.load_model(model, 'complex')
print "done"
# Align all states
cmd.intra_fit('all')
cmd.hide('all')
#cmd.rewind()
cmd.select('receptor', 'not chain C and not hydrogen')
cmd.select('mhc', 'chain A or chain B')
cmd.select('tcr', 'chain D or chain E')
cmd.select('ligand', 'chain C and not hydrogen')
cmd.deselect()
cmd.hide('all')
cmd.show('cartoon', 'receptor')
cmd.show('sticks', 'ligand')
util.cbay('ligand')
cmd.color('green', 'mhc')
cmd.color('red', 'tcr')
from pymol import cmd
from sys import argv
# this script is run automatically by MCsim and will need to be changed for someone else's use
numFrames = int(argv[1])
channel = int(argv[2])
pathPDB = argv[3]
base = int(argv[4])
for idx in range(0, numFrames):
cmd.load(pathPDB + "fine%03dv%i.pdb" % (idx, channel), "snap")
cmd.intra_fit("snap")
cmd.mset("1 -%d" % numFrames)
cmd.show('spheres', 'resn DNA')
#cmd.show('lines', 'resn DNA')
cmd.hide('sticks', 'resn DNA')
cmd.spectrum('count', 'rainbow', 'resn DNA')
if base == 1:
cmd.set("sphere_scale", 0.5)
else:
cmd.set("sphere_scale", 0.15)
#cmd.mplay()
'(x,y,z) = xyz_iter.next()',
space=locals())
#for pdb_index in range(natoms):
#if (pdb_index % 100)==0: print pdb_index
#model_index = model_mapping[pdb_index]
#model.atom[model_index].coord = (10 * ncfile.variables['positions'][iteration, replica, pdb_index, :]).squeeze().tolist()
#for k in range(3):
# model.atom[model_index].coord[k] = float(ncfile.variables['positions'][iteration, replica, pdb_index, k]) * 10.0 # convert to angstroms
#cmd.load_model(model, 'complex', state=iteration+1)
#cmd.load_model(model, 'complex')
print "done"
# Align all states
cmd.intra_fit('all')
cmd.hide('all')
#cmd.rewind()
cmd.select('receptor', '(not resn MOL) and (not resn WAT) and (not hydrogen)')
cmd.select('ligand', 'resn MOL and not hydrogen')
cmd.select('ions', 'resn Na\+ or resn Cl\-')
cmd.deselect()
cmd.hide('all')
cmd.show('spheres', 'receptor')
util.cbaw('receptor')
cmd.show('spheres', 'ligand')
cmd.show('spheres', 'ions')
util.cbay('ligand')
def avgStates(object='all',object_sel=None,first=1,last=0,newobj=None,fitverdict='no',\
verb=0,pairs=1,writefiles=1):
'''
ARGUMENTS:
object (string [defaults to 'all']):
Starting PyMOL molecular object consisting of >1 states to be averaged
object_sel (string [defaults to "name CA"]):
An optional subset of atoms on which to operate (e.g., "name CA and resi 20-50")
first (int [defaults to 1]):
number of the first state to include in averaging
last (int [defaults to last state]):
Number of the last state to include in averaging. A value of '0' (zero)
means use the last state in the object.
newobj (string [defaults to name of object with string "_avg_AtoZ" appended,
where A and Z are the numbers of the first and last frames included
in the averaging]):
Desired name of the new output pymol object (i.e., averaged structure)
fitverdict (string [defaults to "no", any value != "no" is taken as a "yes"]):
Use the pymol function intra_fit() to fit all the states of "object & object_sel"
before calculating the average structure and RMSDs??
'no' = NO, !'no' = yes
verb (int [defaults to 0]):
Verbosity level of output. 0 = quiet, !0 = verbose (e.g., write position-specific
RMSDs to standard output).
pairs (int [default 1]):
Also calculate average pairwise RMSD for each residue position?? 0 = no, !0 = yes
writefiles (int [default 1]):
Write the calculated RMSD data to plaintext files?? 0 = no, !0 = yes
NOTES:
The state specified as 'first' is taken as the reference state for the (optional)
intra_fit() alignment. If no selection is given, the string "and name CA" will be
appended to the object and the averaging and rmsd calculations will be restricted
to "object and name CA" (i.e., C-alpha atoms). To avoid this default behavior,
specify a valid pymol atom selection for the object_sel argument (e.g., "name CA"
or "name CA and resi 20-50" or whatever).
'''
object = str(object)
object_sel = str(object_sel)
first=int(first)
last=int(last)
num_states_tot = cmd.count_states(object)
if last<1:
last=num_states_tot
num_states2avg = last - first + 1
if newobj==None or newobj=='':
newobj = "%s_avg_%dto%d"%(object,first,last)
cmd.create(newobj,object,first,1)
if writefiles:
print '%s'%('-'*80)
datfileprefix = '%s_%dto%d'%(replace(object,' ','_'),first,last)
avg_rmsd_file = datfileprefix + '.avg_rmsd.dat'
avg_file = open(avg_rmsd_file,'w')
print 'Opened "%s" file for writing residue-specific RMSDs to average structure...'%avg_rmsd_file
if pairs:
pair_rmsd_file = datfileprefix + '.pair_rmsd.dat'
pair_file = open(pair_rmsd_file,'w')
print 'Opened "%s" file for writing averaged residue-specific pairwise RMSDs...'%pair_rmsd_file
print '%s'%('-'*80)
else:
avg_file = open('/dev/null','w') # just do this instead of evaluating conditionals each time
if pairs:
pair_file = open('/dev/null','w') # through the atom loops (below)...
obj_sel_string = (lambda o,s: (o and s and "%s and %s"%(o,s)) or \
(o and not s and "%s and name CA"%o))(object,object_sel)
newobj_sel_string = (lambda o,s: (o and s and "%s and %s"%(o,s)) or \
(o and not s and "%s and name CA"%o))(newobj,object_sel)
print '%s'%('-'*80)
print 'Averaging %d states [%d, %d] of object "%s" (%d total states) to get new single-state object "%s"...' \
% (num_states2avg,first,last,obj_sel_string,num_states_tot,newobj)
print '%s'%('-'*80)
tmpobject='%s_tmp4avg_%dto%d'%(object,first,last)
i=0
for eachstate in range(first,last+1):
i+=1 # because pymol states are indexed starting from 1 (not 0)
cmd.create(tmpobject,obj_sel_string,eachstate,i)
if fitverdict != 'no':
print "-> proceeding WITH FITTING to reference state..."
tmpobject_intrafit_rmsds = cmd.intra_fit(tmpobject,1)
if verb:
print ' intrafit_rmsds = ',
print tmpobject_intrafit_rmsds
else:
print "-> proceeding WITHOUT FITTING to reference state..."
# create an atom index map between original (complete) object and subset to be averaged:
atindex_map = [None]
for at in cmd.get_model(newobj_sel_string,1).atom:
atindex_map.append(at.index)
if verb:
print "-> atom index_map = ",
print atindex_map
newobj_chempy = cmd.get_model(tmpobject,1)
for at in newobj_chempy.atom:
this_at_idx = at.index
sum_x = sum_y = sum_z = 0.0
avg_x = avg_y = avg_z = 0.0
state_coords=[]
rmsd_sum = rmsd = 0.0
for s in range(1,num_states2avg+1):
this_x = cmd.get_model(tmpobject,s).atom[this_at_idx-1].coord[0]
this_y = cmd.get_model(tmpobject,s).atom[this_at_idx-1].coord[1]
this_z = cmd.get_model(tmpobject,s).atom[this_at_idx-1].coord[2]
sum_x += this_x
sum_y += this_y
sum_z += this_z
state_coords.append([this_x,this_y,this_z])
avg_x = sum_x / num_states2avg
avg_y = sum_y / num_states2avg
avg_z = sum_z / num_states2avg
cmd.alter_state(1,'%s and id %d'%(newobj,atindex_map[this_at_idx]),'x=%f'%avg_x)
cmd.alter_state(1,'%s and id %d'%(newobj,atindex_map[this_at_idx]),'y=%f'%avg_y)
cmd.alter_state(1,'%s and id %d'%(newobj,atindex_map[this_at_idx]),'z=%f'%avg_z)
# position-specific RMSDs with respect to average structure:
for somept in state_coords:
rmsd_sum += pow(calcDist(somept,[avg_x,avg_y,avg_z]),2.0)
rmsd = sqrt(rmsd_sum / num_states2avg)
# DEBUG:
# DEBUG print 'newobj = %s / id = %d / b = %0.3f'%(newobj,atindex_map[this_at_idx],rmsd)
cmd.alter('%s and id %d'%(newobj,atindex_map[this_at_idx]),'b=%0.3f'%rmsd)
if verb:
print '"%s" index = %d'%(obj_sel_string,this_at_idx)
print 'rmsd = %0.3f'%rmsd
avg_file.write('%d\t%0.3f\n'%(atindex_map[this_at_idx],rmsd))
# position-specific RMSDs averaged pairwise over the bundle:
if pairs:
for at in newobj_chempy.atom:
this_at_idx = at.index
loop_counter = 0
running_sum = 0.0
for s1 in range(1,num_states2avg+1):
for s2 in range(s1+1,num_states2avg+1):
#DB print 'computing position %d, %d x %d'%(this_at_idx,s1,s2)
pos1 = [ cmd.get_model(tmpobject,s1).atom[this_at_idx-1].coord[0],
cmd.get_model(tmpobject,s1).atom[this_at_idx-1].coord[1],
cmd.get_model(tmpobject,s1).atom[this_at_idx-1].coord[2] ]
pos2 = [ cmd.get_model(tmpobject,s2).atom[this_at_idx-1].coord[0],
cmd.get_model(tmpobject,s2).atom[this_at_idx-1].coord[1],
cmd.get_model(tmpobject,s2).atom[this_at_idx-1].coord[2] ]
loop_counter += 1
running_sum += pow(calcDist(pos1,pos2),2.0)
# print '%s'%('.'*s2)
pairwise_rmsd = sqrt(running_sum / loop_counter)
if verb: print 'calculated pairwise RMSD for %d pairs at position %d = %0.3f' % \
(loop_counter,atindex_map[this_at_idx],pairwise_rmsd)
pair_file.write('%d\t%0.3f\n'%(atindex_map[this_at_idx],pairwise_rmsd))
if writefiles:
avg_file.close()
if pairs:
pair_file.close()