def morph_movie(morph,
view,
color,
base_name,
frameno_offset=0,
morph_reverse=False):
number_of_states = cmd.count_states(morph)
# after this I will have each state available as its own object, called "{morphname}_000x" (padded to length of 4)
# x starts with 1 --> not clear - that might depend on the numbering in the input file
cmd.split_states(morph)
last_frame = frameno_offset
for statenum in range(0, number_of_states + 1):
if morph_reverse:
statenum = max(1, cmd.count_states(morph) - statenum)
else:
statenum = min(statenum + 1, cmd.count_states(morph))
state_name = morph + "_" + str(statenum).zfill(4)
clump_representation([state_name], color, state_name),
cmd.set_view(view)
cmd.png(base_name + str(last_frame).zfill(3),
width=1920,
height=1080,
ray=True)
clump_cleanup([state_name], state_name)
cmd.remove(state_name)
last_frame += 1
return last_frame
def mican(mobile, target, option=""):
'''
DESCRIPTION
Align two (ligand) selections based on mican algorithm.
ARGUMENTS
mobile = str: atom selection of mobile object
target = str: atom selection of target object
EXAMPLE
fetch 3zcf 4n8t, bsync=0
mcsalign /3zcf//A/HEC, /4n8t//A/HEM
zoom /4n8t//A/HEM, animate=2, buffer=3
'''
import subprocess
import tempfile
import os
with tempfile.TemporaryDirectory() as dname:
# print tmp dir name
print("Temporary directory =" + dname)
# make sure you have mican in PATH
# directly giving 'execute' full path below is good alternative
# For example : execute = "/usr/bin/mican"
execute = "mican"
tmptarget = dname + "/target.pdb"
tmpmobile = dname + "/mobile.pdb"
tmpout = dname + "/aligned.pdb"
# save pdb for mican
cmd.save(tmptarget, target)
cmd.save(tmpmobile, mobile)
modeoption = "-" + option
option2 = "-o"
outfile = tmpout
mican = [execute, tmpmobile, tmptarget, option2, outfile]
for op in option.split():
if(op == "-o"):
print("option -o is reserved")
raise CmdException
mican.append(op)
proc=subprocess.run(mican,stdout = subprocess.PIPE)
print(proc.stdout.decode("utf8")) # print result to pymol console
cmd.load(outfile, "aligned")
cmd.split_states("aligned")
cmd.select("mobileback",mobile + " and backbone")
cmd.align("mobileback", "aligned_0001 and backbone")
# use cmd pair_fit if you think align is not good
# print("Using cmd.align instead of cmd.pair_fit")
# pymol.cmd.pair_fit("mobileback", "aligned_0001 and backbone")
cmd.delete("mobileback")
cmd.delete("aligned")
cmd.delete("aligned_0001")
cmd.delete("aligned_0002")
def export_pymol(file1, file2):
py_object1 = file1.rsplit(".sdf", maxsplit=1)[0]
py_object2 = file2.rsplit(".sdf", maxsplit=1)[0]
pref1 = py_object1.split("/")[-1]
pref2 = py_object2.split("/")[-1]
cmd.load(filename=file1)
cmd.set_name(pref1, "query_conf")
cmd.load(filename=file2)
cmd.split_states(object="query_conf")
cmd.split_states(object=pref2)
cmd.delete("query_conf")
cmd.delete(pref2)
cmd.save(f"{py_object2}.pse")
cmd.delete("all")
def split_states_chains_save( sel1 ):
'''
Simple script to split states of NMR ensemble and then save the complex and chain A separately for each state
'''
cmd.remove("resn hoh")
for ind in range(1,cmd.count_states(sel1)+1):
if cmd.count_states(sel1) > 1:
cmd.split_states(sel1)
file_prefix="{0}_{1:04d}".format(sel1, ind)
else:
file_prefix=sel1
cmd.save(file_prefix + "_complex.pdb", file_prefix)
cmd.split_chains(file_prefix)
cmd.save(file_prefix + ".pdb", file_prefix + "_A")
def get_ligand(ligand_path):
'''
DESCRIPTION
Reads in docked poses generated by Vina, saves first mode (highest binding affinity) as ligand.mol2
'''
base = os.path.basename(ligand_path)
FILENAME = os.path.splitext(base)[0]
#print "Loading %s\n", % (ligand_path)
cmd.load(
ligand_path
) # load "C:/Users/Janet Liu/Documents/2019-2020/CS221/Project/data_files/stx_hnav1-7/ligand_out.pdbqt"
cmd.split_states(FILENAME)
cmd.save(
os.path.dirname(ligand_path) + "/ligand.mol2", FILENAME + "_0001", -1,
'mol2')
def toGroup(groupName, sel, prefix="", delOrig=True):
"""
DESCRIPTION
toGroup will take a multistate object and extract it
to a group with N objects all in state #1. It essentially
performs the following:
split_states myObj, prefix=somePrefix
group newGroup, somePrefix*
delete myObj
PARAMETERS:
groupName (string)
The name of the group to create
sel (string)
The name of the selection/object from which
to make the group
prefix (string)
The prefix of the names of each of split states.
For example, if your prefix is ''obj'' and is in
states 1 through 100 then the states will be labeled
obj1, obj2, obj3, ..., obj100.
delOrig (string/boolean)
If true then delete the original selection, otherwise not.
RETURN
Nothing, it makes a new group.
"""
if prefix == "":
prefix = sel + "_grouped"
cmd.split_states(sel, prefix=prefix)
cmd.group(groupName, prefix + "*")
if delOrig:
cmd.delete(sel)
def toGroup(groupName,sel,prefix="",delOrig=True):
"""
DESCRIPTION
toGroup will take a multistate object and extract it
to a group with N objects all in state #1. It essentially
performs the following:
split_states myObj, prefix=somePrefix
group newGroup, somePrefix*
delete myObj
PARAMETERS:
groupName (string)
The name of the group to create
sel (string)
The name of the selection/object from which
to make the group
prefix (string)
The prefix of the names of each of split states.
For example, if your prefix is ''obj'' and is in
states 1 through 100 then the states will be labeled
obj1, obj2, obj3, ..., obj100.
delOrig (string/boolean)
If true then delete the original selection, otherwise not.
RETURN
Nothing, it makes a new group.
"""
if prefix=="":
prefix=sel + "_grouped"
cmd.split_states(sel, prefix=prefix)
cmd.group(groupName,prefix+"*")
if delOrig:
cmd.delete(sel)
def extract_state_to_object(morph, state, new_object):
number_of_states = cmd.count_states(morph)
if number_of_states < state:
print("in extract_state_to_object() requested state " +
"{} > number of states ({}) available in {}".format(
state, number_of_states, morph))
exit()
if morph == new_object:
print("in extract_state_to_object() please picked different " +
"name for extraction (currently bouth '{}') ".format(morph))
exit()
# after this I will have each state available as its own object, called "{morphname}_000x" (padded to length of 4)
# x starts with 1 --> not clear - that might depend on the numbering in the input file
cmd.split_states(morph)
state_name = morph + "_" + str(state).zfill(4)
cmd.copy(new_object, state_name)
for statenum in range(1, number_of_states + 1):
state_name = morph + "_" + str(statenum).zfill(4)
cmd.delete(state_name)
cmd.delete(morph)
return
def test_split_states(self):
cmd.fragment('ala', 'm1')
cmd.create('m1', 'm1', 1, 2)
cmd.create('m1', 'm1', 1, 3)
cmd.split_states('m1', 1, 2)
self.assertItemsEqual(['m1', 'm1_0001', 'm1_0002'], cmd.get_names())
def flatten_obj(name="",selection="",state=0,rename=0,quiet=1,chain_map=""):
"""
DESCRIPTION
"flatten_obj" combines multiple objects or states into a single object,
renaming chains where required
USAGE
flatten_obj name, selection[, state[, rename[, quiet[, chain_map]]]]
ARGUMENTS
name = a unique name for the flattened object {default: flat}
selection = the set of objects to include in the flattening. The selection
will be expanded to include all atoms of objects. {default: all}
state = the source state to select. Use 0 or -1 to flatten all states {default: 0}
rename = The scheme to use for renaming chains: {default: 0}
(0) preserve chains IDs where possible, rename other chains
alphabetically
(1) rename all chains alphabetically
(2) rename chains using the original chain letter, object name, and state
quiet = If set to 0, print some additional information about progress and
chain renaming {default: 1}
chain_map = An attribute name for the 'stored' scratch object. If
specified, `stored.<chain_map>` will be populated with a dictionary
mapping the new chain names to a tuple giving the originated object,
state, and chainID. {default: ""}
NOTES
Like the select command, if name is omitted then the default object name
("flat") is used as the name argument.
Chain renaming is tricky. PDB files originally limited chains to single
letter identifiers containing [A-Za-z0-9]. When this was found to be
limiting, multi-letter chains (ideally < 4 chars) were allowed. This is
supported as of PyMOL 1.7. Earlier versions do not accept rename=2, and
will raise an exception when flattening a structure with more than 62
chains.
EXAMPLES
flatten_obj flat, nmrObj
flatten_obj ( obj1 or obj2 )
SEE ALSO
split_states
"""
# arguments
# Single argument; treat as selection
if name and not selection:
selection = name
name = ""
# default name and selection
if not name:
name = "flat"
if not selection:
selection = "(all)"
state = int(state)
rename = int(rename)
quiet = int(quiet)
# Wrap in extra parantheses for get_object_list
selection = "( %s )" % selection
if rename == 0:
chainSet = DefaultChainSet()
elif rename == 1:
chainSet = SequentialChainSet()
elif rename == 2:
chainSet = LongChainSet()
else:
raise ValueError("Unrecognized rename option (Valid: 0,1,2)")
metaprefix = "temp" #TODO unique prefix
# store original value of retain_order, which causes weird interleaving of
# structures if enabled.
retain_order = cmd.get("retain_order")
try:
cmd.set("retain_order",0)
# create new object for each state
for obj in cmd.get_object_list(selection):
if state <= 0:
# all states
prefix = "%s_%s_"%(metaprefix,obj)
cmd.split_states(obj,prefix=prefix)
else:
prefix = "%s_%s_%04d"%(metaprefix,obj,state)
cmd.create(prefix, obj, state, 1)
# renumber all states
statere = re.compile("^%s_(.*)_(\d+)$" % metaprefix) # matches split object names
warn_lowercase = False
# Iterate over all objects with metaprefix
try:
for obj in cmd.get_object_list("(%s_*)"%(metaprefix) ):
m = statere.match(obj)
if m is None:
print(("Failed to match object %s" %obj))
continue
origobj = m.group(1)
statenum = int(m.group(2))
chains = cmd.get_chains(obj)
rev_chain_map = {} #old -> new, for this obj only
for chain in sorted(chains,key=lambda x:(len(x),x)):
new_chain = chainSet.map_chain(origobj,statenum,chain)
rev_chain_map[chain] = new_chain
if not quiet:
print((" %s state %d chain %s -> %s"%(origobj,statenum,chain, new_chain) ))
if not _long_chains:
if len(new_chain) > 1:
raise OutOfChainsError("No additional chains available (max 62).")
space = {'rev_chain_map':rev_chain_map}
cmd.alter(obj,"chain = rev_chain_map[chain]",space=space)
print(("Creating object from %s_*"%metaprefix))
# Recombine into a single object
cmd.create(name,"%s_*"%metaprefix)
# Set chain_map
if chain_map:
setattr(stored,chain_map,chainSet)
# Warn if lowercase chains were generated
if cmd.get("ignore_case") == "on" and any([c.upper() != c for c in list(chainSet.keys())]):
print("Warning: using lower-case chain IDs. Consider running the "
"following command:\n set ignore_case, 0" )
finally:
# Clean up
print("Cleaning up intermediates")
cmd.delete("%s_*"%metaprefix)
finally:
# restore original parameters
print("Resetting variables")
cmd.set("retain_order",retain_order)
def scene_interpolate(view_init_str,
object_properties,
base_name,
number_of_frames=15,
frameno_offset=0,
view_last_str=None,
morph_properties=None):
# the hope is to get rid of this some day
# though it does not seem that it will happen before we move to blender
global lipid_resnums
if "lipid" in object_properties.keys():
style_lipid("lipid")
props = object_properties["lipid"]
# the fifth argument is transparency - so the attempt was made to fiddle with lipd transparency
if len(props) > 5:
# lipid is a special problem because I chose to represent it with sticks and transparency
# does not work for raytraced sticks (duh)
with open("{}/{}".format(structure_home,
structure_filename["lipid"])) as inf:
for line in inf:
if line[:4] != "ATOM": continue
lipid_resnums.add(line.split()[4])
if morph_properties:
morph_lengths = set()
for morph_name in morph_properties.keys():
morph_lengths.add(cmd.count_states(morph_name))
cmd.split_states(morph_name)
if len(morph_lengths) > 1:
print("morphs are all expected to be of the same length; found",
morph_lengths)
exit()
if len(
morph_lengths
) == 0: # can this happen? morph_properties should be False in thate casem shouldn't it
morph_properties = None
else:
number_of_frames = morph_lengths.pop()
# I could not find the way to move the surface, if there's one provided
# so I'm moving the object and re-creating the mesh
view_init = view_string2view(view_init_str) if view_last_str else None
view_last = view_string2view(view_last_str) if view_last_str else None
# get quaternions for interpolation
qstart = view2quat(view_init) if view_last_str else None
qend = view2quat(view_last) if view_last_str else None
for objnm in object_properties.keys():
cmd.hide("everything", objnm)
clump_cleanup([objnm], objnm)
last_frame = frameno_offset
for frameno in range(0, number_of_frames + 1):
# object position interpolation
tmpnames = object_tfm_interpolate(object_properties, number_of_frames,
frameno)
# view interpolation, if called for
if not view_last_str:
view = view_init_str
else:
view = view2view_string(
intermediate_view(view_init, view_last, qstart, qend,
number_of_frames, frameno))
# morph
if morph_properties:
tmp_obj_props = {}
for morph_name, props in morph_properties.items():
[
morph_color, morph_reverse, tfm_from, tf_to, tfm_reverse,
transparency
] = props
# change shape
if morph_reverse:
stateno = max(1, cmd.count_states(morph_name) - frameno)
else:
stateno = min(frameno + 1, cmd.count_states(morph_name))
morph_state_name = morph_name + "_" + str(stateno).zfill(4)
# [identity_tfm, gbg_tfm, tfm_reverse, color, small_molecule rep]
tmp_obj_props[morph_state_name] = [
tfm_from, tf_to, tfm_reverse, morph_color, False,
transparency
]
# reposition all morphs and show them as clumps
tmpnames.extend(
object_tfm_interpolate(tmp_obj_props, number_of_frames,
frameno))
# some of the ops above move camera (why the f do you have to move the camera to create a new rep?)
cmd.set_view(view)
cmd.png(base_name + str(last_frame).zfill(3),
width=1920,
height=1080,
ray=True)
object_cleanup(tmpnames)
last_frame += 1
return last_frame
def test_split_states(self):
cmd.fragment('ala', 'm1')
cmd.create('m1', 'm1', 1, 2)
cmd.create('m1', 'm1', 1, 3)
cmd.split_states('m1', 1, 2)
self.assertItemsEqual(['m1', 'm1_0001', 'm1_0002'], cmd.get_names())
def visual(self, uno):
cmd.reinitialize()
nombre = uno.replace(".dlg", ".pdb")
cmd.load(self.dirct + nombre, "COMP")
cmd.split_states("COMP")
# sys.path.insert(1, '/usr/local/Cellar/pymol/1.7.0.0/lib/python2.7/site-packages/')
import pymol
from pymol import cmd
pymol.finish_launching()
if len(sys.argv) != 6:
print("Usage: %s pdb-file align-sele save-sele align-atoms frame-pdb")
sys.exit(1)
struct_path = sys.argv[1]
struct_name = os.path.splitext(os.path.basename(struct_path))[0]
cmd.load(struct_path, struct_name)
cmd.split_states(struct_name, 1, 1, "temp")
cmd.delete(struct_name)
# cmd.save('%s_debug.pse' % struct_name)
cmd.set_name("temp0001", struct_name)
frame_bname = "frame"
sys
# frame_path = './aa_frames/ile_frame.pdb'
frame_path = sys.argv[5]
cmd.load(frame_path, frame_bname)
# target_sele = 'resn asp'
target_sele = sys.argv[2]
# dist = 6
# dist = sys.argv[3]
user_sele = sys.argv[3]
def split_alignment(object, object1=None, object2=None, delimiter="_"):
"""DESCRIPTION
Splits a two-state object into two separate objects.
USAGE
split_alignment object[, object1, object2][, delimiter]
ARGUMENTS
object Two-state input object
object1 What to name the first state from the object [optional]
object2 What to name the second state from the object [optional]
delimiter Delimiter which separates object1 and object2 in the object name.
See DETAILS. [default '_']
DETAILS
The input object must contain at least two states. Additional states are
ignored.
If object1 and object2 are ommitted, the script will attempt to generate
them based on the input object's name. The name is split around the first
instance of <delimiter>. Thus, objects which follow the convention
"object1_object2" will be properly split without additional arguments.
EXAMPLES
# Results in objects '1AX8.A' and '3PIV.A'
split_alignment 1AX8.A_3PIV.A
# Results in objects 'query' and 'target'
split_alignment alignment, query, target
# Results in objects '1AX8.A' and '3PIV.A'
split_alignment 1AX8.A_vs_3PIV.A, delimiter="_vs_"
"""
# check that we have at least two states
if cmd.count_states(object) < 2:
print ("Error: input object must contain at least two states.")
return
prefix = "split%04d_" % random.randint(0, 9999) # make unique
# guess output names
if object1 is None and object2 is None:
try:
d = object.index(delimiter)
object1 = object[:d]
object2 = object[d + len(delimiter) :]
except:
object1 = "%s_%04d" % (object, 1)
object2 = "%s_%04d" % (object, 2)
print "Warning: '%s' not found in '%s'. Using names %s and %s." % (delimiter, object, object1, object2)
# split them
cmd.split_states(object, prefix=prefix)
# rename to output names
cmd.set_name("%s%04d" % (prefix, 1), object1)
cmd.set_name("%s%04d" % (prefix, 2), object2)
# delete other states
for o in cmd.get_names("objects"):
if o.startswith(prefix):
cmd.delete(o)
def flatten_obj(name="",
selection="",
state=0,
rename=0,
quiet=1,
chain_map=""):
"""
DESCRIPTION
"flatten_obj" combines multiple objects or states into a single object,
renaming chains where required
USAGE
flatten_obj name, selection[, state[, rename[, quiet[, chain_map]]]]
ARGUMENTS
name = a unique name for the flattened object {default: flat}
selection = the set of objects to include in the flattening. The selection
will be expanded to include all atoms of objects. {default: all}
state = the source state to select. Use 0 or -1 to flatten all states {default: 0}
rename = The scheme to use for renaming chains: {default: 0}
(0) preserve chains IDs where possible, rename other chains
alphabetically
(1) rename all chains alphabetically
(2) rename chains using the original chain letter, object name, and state
quiet = If set to 0, print some additional information about progress and
chain renaming {default: 1}
chain_map = An attribute name for the 'stored' scratch object. If
specified, `stored.<chain_map>` will be populated with a dictionary
mapping the new chain names to a tuple giving the originated object,
state, and chainID. {default: ""}
NOTES
Like the select command, if name is omitted then the default object name
("flat") is used as the name argument.
Chain renaming is tricky. PDB files originally limited chains to single
letter identifiers containing [A-Za-z0-9]. When this was found to be
limiting, multi-letter chains (ideally < 4 chars) were allowed. This is
supported as of PyMOL 1.7. Earlier versions do not accept rename=2, and
will raise an exception when flattening a structure with more than 62
chains.
EXAMPLES
flatten_obj flat, nmrObj
flatten_obj ( obj1 or obj2 )
SEE ALSO
split_states
"""
# arguments
# Single argument; treat as selection
if name and not selection:
selection = name
name = ""
# default name and selection
if not name:
name = "flat"
if not selection:
selection = "(all)"
state = int(state)
rename = int(rename)
quiet = int(quiet)
# Wrap in extra parantheses for get_object_list
selection = "( %s )" % selection
if rename == 0:
chainSet = DefaultChainSet()
elif rename == 1:
chainSet = SequentialChainSet()
elif rename == 2:
chainSet = LongChainSet()
else:
raise ValueError("Unrecognized rename option (Valid: 0,1,2)")
metaprefix = "temp" #TODO unique prefix
# store original value of retain_order, which causes weird interleaving of
# structures if enabled.
retain_order = cmd.get("retain_order")
try:
cmd.set("retain_order", 0)
# create new object for each state
for obj in cmd.get_object_list(selection):
if state <= 0:
# all states
prefix = "%s_%s_" % (metaprefix, obj)
cmd.split_states(obj, prefix=prefix)
else:
prefix = "%s_%s_%04d" % (metaprefix, obj, state)
cmd.create(prefix, obj, state, 1)
# renumber all states
statere = re.compile("^%s_(.*)_(\d+)$" %
metaprefix) # matches split object names
warn_lowercase = False
# Iterate over all objects with metaprefix
try:
for obj in cmd.get_object_list("(%s_*)" % (metaprefix)):
m = statere.match(obj)
if m is None:
print(("Failed to match object %s" % obj))
continue
origobj = m.group(1)
statenum = int(m.group(2))
chains = cmd.get_chains(obj)
rev_chain_map = {} #old -> new, for this obj only
for chain in sorted(chains, key=lambda x: (len(x), x)):
new_chain = chainSet.map_chain(origobj, statenum, chain)
rev_chain_map[chain] = new_chain
if not quiet:
print((" %s state %d chain %s -> %s" %
(origobj, statenum, chain, new_chain)))
if not _long_chains:
if len(new_chain) > 1:
raise OutOfChainsError(
"No additional chains available (max 62).")
space = {'rev_chain_map': rev_chain_map}
cmd.alter(obj, "chain = rev_chain_map[chain]", space=space)
print(("Creating object from %s_*" % metaprefix))
# Recombine into a single object
cmd.create(name, "%s_*" % metaprefix)
# Set chain_map
if chain_map:
setattr(stored, chain_map, chainSet)
# Warn if lowercase chains were generated
if cmd.get("ignore_case") == "on" and any(
[c.upper() != c for c in list(chainSet.keys())]):
print(
"Warning: using lower-case chain IDs. Consider running the "
"following command:\n set ignore_case, 0")
finally:
# Clean up
print("Cleaning up intermediates")
cmd.delete("%s_*" % metaprefix)
finally:
# restore original parameters
print("Resetting variables")
cmd.set("retain_order", retain_order)
'--hetatm',
help=
'force all the atom to be defined as HETATM type. The HETATM atom have CONECT defined by default.',
action='store_true')
parser.add_argument('--addh', help='Add hydrogens', action='store_true')
parser.add_argument('--select',
help='Keep only the selected atoms',
default='all')
args = parser.parse_args()
cmd.load(args.inp, 'inmol')
if args.addh:
cmd.h_add('all')
if args.hetatm:
cmd.alter('all', 'type="HETATM"')
cmd.split_states('inmol')
cmd.remove('inmol')
basename = os.path.basename(os.path.splitext(args.inp)[0])
if args.format is None:
outfmt = os.path.splitext(args.inp)[1][1:]
else:
outfmt = args.format
all_objects = cmd.get_object_list()
try:
os.mkdir(args.outdir)
except FileExistsError:
pass
for i, obj in enumerate(all_objects):
sys.stdout.write(f'{i+1}/{len(all_objects)} saving {obj}\r')
cmd.save(f'{args.outdir}/{basename}_{i:04d}.{outfmt}',
f'{obj} and {args.select}')
if not os.path.exists(pdbpath):
pdbpath = os.path.join(pbdir, 'final.pdb')
if not os.path.exists(pdbpath) or not os.path.exists(cypath):
raise SystemExit # does not seem like a ponderosa result directory
f = open(cypath, 'r')
ordered = f.read()
f.close()
oreg = ordered.replace(',', '+') # to alleviate a small bug.
ordered = oreg.replace('+', ', ') # more common in papers
from pymol import cmd
cmd.load(pdbpath, 'for_rmsd')
cmd.split_states('for_rmsd')
# Backbone
bb_rmsds = []
for i in range(2, 21):
fit_result = cmd.align('for_rmsd_0001 & i. %s & n. N+CA+C+O' % (oreg),
'for_rmsd_%04d' % (i),
cutoff=0)
bb_rmsds += [
fit_result[0],
]
# All heavy atoms
ha_rmsds = []
for i in range(2, 21):
fit_result = cmd.align('for_rmsd_0001 & i. %s & n. N*+C*+O*+S*' % (oreg),
in_msg = 'Select the sampling density from 1-4. \n\
Higher density means higher accuracy but slower performance.'
density = int(s.show_inputdialog('Select the Sampling density', in_msg, '2'))
if density < 1 or density > 4:
s.show_message('Error', 'The sampling density is invalid.')
raise SystemExit
in_msg = 'Select the model number. \n\
Use 1 if there\'s only one model. Use 0 for all models.'
model = int(s.show_inputdialog('Select the model', in_msg, '1'))
cmd.delete('all')
cmd.load(pdbpath, 'for_area')
cmd.split_states('for_area')
cmd.set('dot_solvent', 1)
cmd.set('dot_density', density)
if model != 0: # for a specific model
try:
area = cmd.get_area('resi ' + selection +
' and model for_area_' + '{:04d}'.format(model))
except:
s.show_message('Error', 'The model number was wrong.')
raise SystemExit
print(area)
message = 'The surface area for residue ' + selection + ' is ' + \
'{:.3f}'.format(area) + ' Angstroms^2'
from pymol import cmd
########### CHANGE THIS LIST AS NEEDED #######
delete_atoms = ['N21', 'C26', 'H27', 'H28',
'H29'] # atoms to be removed from rotamer library
three_letter_code = 'NIR' # three-letter residue name specified in -n for molfile_to_params
atoms_to_align = [
'CL', 'CL1', 'CL2'
] # atoms around the lysine-ligand bond for alignment and visualisation
##############################################
cmd.delete('*')
cmd.load('conformers.mol2')
target = 'conformers'
cmd.split_states(target, prefix="Molecule_Name_")
#cmd.split_states(target)
cmd.delete(target)
counter = 0
#target_atom1=''
#target_atom2=''
#target_atom3=''
#target_atom4=''
for object in cmd.get_object_list('(all)'):
cmd.alter(object, 'resn="' + three_letter_code + '"')
counter += 1
if counter == 1:
target_object = object
target_atom1 = target_object + ' and name ' + atoms_to_align[0]
target_atom2 = target_object + ' and name ' + atoms_to_align[1]
target_atom3 = target_object + ' and name ' + atoms_to_align[2]
def mutate(
self, pdbid: str,
replace_with: Dict[int,
Optional[str]]) -> Union[List[str], rdchem.Mol]:
"""Modify amino acid residues at the defined positions.
If the locations indexes exceed the amount in data, then they
will be ignored and a warning will be announced.
Params:
-------
pdbid: str
PDB ID associated with the structure.
replace_with: dict
The index location(s) within the full protein to replace
certain residue(s) with. If a residue associated with a
index location is None, then the modified residue is chosen
randomly. If a certain index exceeds the number of available
residues in the protein, then those enteries are simply
ignored and the user is notified.
Returns:
--------
protein: list of str or rdkit.Chem.rdchem.Mol
Modified protein with residues. If fmt="primary", then list
of string (peptide names) is returned. If fmt="tertiary",
then 3D molecule structure is returned.
"""
# Load PDB structure (download, if necessary)
pdb_dir = maybe_create_dir(os.path.join(self.rootdir, pdbid))
pdb_file = os.path.join(pdb_dir, f"{pdbid}.pdb")
if not os.path.exists(pdb_file):
is_successful = cmd.fetch(pdbid,
name=pdbid,
state=1,
type="pdb",
path=pdb_dir)
if is_successful == -1:
raise DownloadError(f"Unable to download '{pdbid}'.")
else:
cmd.load(pdb_file, object=pdbid, state=1, format="pdb")
# Get all residue names, see: https://pymolwiki.org/index.php/List_Selection
resnames_dict = {"names": []}
cmd.iterate("(name ca)", "names.append(resn)", space=resnames_dict)
residue_names = resnames_dict["names"]
num_residues = len(residue_names)
# Cleanup idxs: remove indicies that exceed number of available residues
nonvalid_idxs = [
idx for idx in replace_with.keys() if idx > num_residues
]
for idx in nonvalid_idxs:
print(
f"OutOfRange: Removing idx {idx} (only {num_residues} residues)."
)
replace_with.pop(idx)
# Randomly choose an amino acid (AA) to replace residue, if None is provided.
# Additionally, format string such that it is a valid 3 letter amino acid.
for idx, residue in replace_with.items():
if residue is None:
replace_with[idx] = np.random.choice(aa3)
elif is_aa(residue):
residue = residue.upper()
if len(residue) == 1:
replace_with[idx] = one_to_three.get(residue)
elif len(residue) == 3:
replace_with[idx] = residue
else:
raise ValueError(
f"Invalid residue '{residue}'. Choose one from "
f"the following {aa1+aa3}.")
# Determine save filepath name
modified_res_str = ":".join(
[f"{k}{three_to_one.get(v)}" for k, v in replace_with.items()])
filename = f"{self.fmt}_{modified_res_str}"
filename += ".pdb" if self.fmt == "tertiary" else ".json"
save_filepath = os.path.join(self.rootdir, pdbid, filename)
# Replace primary structure, i.e. residue names (str)
if self.fmt == "primary":
# Load data from cache, if it exists
protein = None
if os.path.exists(save_filepath):
with open(save_filepath) as json_file:
protein = json.load(json_file)
if protein is None:
for idx, residue in replace_with.items():
residue_names[
idx - 1] = residue # since PDB starts with idx of 1
protein = [three_to_one.get(name) for name in residue_names]
# Save sequence temporarily
_ = maybe_create_dir(save_filepath)
with open(save_filepath, "w") as outfile:
json.dump(protein, outfile)
# Replace tertiary structure, i.e. residue's 3D coordinates
elif self.fmt == "tertiary":
if not os.path.exists(save_filepath):
# Split states so that we can optimize only on specific state(s).
# NOTE: Might be useful to choose lowest energy state to mutate,
# OR mutate rotamers for all positions, then choose one with
# lowest energy.
cmd.split_states(object=pdbid)
# Delete all other objects other than one we want to mutate
# NOTE: For now, keep only first object. This might change
# depending on which state needs to be kept.
objs = cmd.get_object_list() # aka states
keep_objs = [pdbid + "_0001"]
for obj in objs:
if obj not in keep_objs:
cmd.delete(obj)
assert keep_objs == cmd.get_object_list()
# Mutate residues
cmd.wizard("mutagenesis")
wizard: Mutagenesis = cmd.get_wizard()
for idx, res in replace_with.items():
selection = "{0:s}//A/{1:d}/".format(keep_objs[0], idx)
wizard.do_select(
selection) # select which residue index to replace
wizard.set_mode(
res) # choose name of residue to replace with
wizard.do_state(
1
) # select rotamer with least strain (aka conflicts w/ other atoms)
wizard.apply() # apply point mutation
cmd.set_wizard(None) # close wizard
# Save PDB temporarily
_ = maybe_create_dir(save_filepath)
cmd.save(save_filepath, selection=pdbid, format="pdb")
cmd.delete("all") # remove all objects, clears workspace
# Load + choose model/structure with lowest energy
# NOTE: If sanitize=True, the function checks if Mol has the correct
# hybridization/valance structure (aka is it chemically reasonable).
# When converting from the PDB block, this sometimes results in
# improper parsing. Instead, for now, we just check if the Mol is
# syntactically valid (i.e. all rings/branches closed, no illegal
# atom types, etc).
protein = rdmolfiles.MolFromPDBFile(save_filepath,
sanitize=False,
removeHs=False)
if protein.GetNumConformers() > 1:
protein = _get_conformer(protein, conformer="min", algo="MMFF")
else:
raise NotImplementedError
# Remove file, if not needed
if not self.cache:
os.remove(save_filepath)
return protein