def add_com_to_pdb(mhc_com, vtcr_com, sample_structure):
"""
Function to add pseudoatoms at MHC-CoM, TCR-CoM, and XYZ axise to the output PDB file
"""
# mhc_com
mhc_com_chain = "X"
sample_structure.add(Chain.Chain(mhc_com_chain))
res_id = (" ", 1, " ")
new_residue = Residue.Residue(res_id, "MCM", " ")
new_atom = Atom.Atom("C", mhc_com, 0, 0.0, " ", "C", 1, "C")
new_residue.add(new_atom)
sample_structure.child_dict[mhc_com_chain].add(new_residue)
# tcr com
tcr_com_chain = "Y"
sample_structure.add(Chain.Chain(tcr_com_chain))
res_id = (" ", 1, " ")
new_residue = Residue.Residue(res_id, "TCM", " ")
new_atom = Atom.Atom("C", vtcr_com, 0, 0.0, " ", "C", 1, "C")
new_residue.add(new_atom)
sample_structure.child_dict[tcr_com_chain].add(new_residue)
# X,Y,Z atoms
pos = [[50, 0, 0], [0, 50, 0], [0, 0, 50]]
resn = ["X", "Y", "Z"]
xyz_chain = "Z"
sample_structure.add(Chain.Chain(xyz_chain))
for i in [0, 1, 2]:
res_id = (" ", i + 1, " ")
new_residue = Residue.Residue(res_id, resn[i], " ")
new_atom = Atom.Atom("O", pos[i], 0, 0.0, " ", "O", 1, "O")
new_residue.add(new_atom)
sample_structure.child_dict[xyz_chain].add(new_residue)
return sample_structure
def get_het_id(hetname, chain: Chain):
residues = chain.get_residues()
for residue in residues:
residue_id = residue.get_id()
hetfield = residue_id[0]
if hetfield != " " and hetfield != "W":
if residue_id[0] == hetname:
return residue_id
else:
return None
def get_het_ids(chain: Chain):
residues = chain.get_residues()
residue_list = []
for residue in residues:
residue_id = residue.get_id()
het_field = residue_id[0]
het_name = residue.get_resname()
if het_field != " " and het_field != "W":
residue_list.append(het_name)
return residue_list
def get_structure_slice_by_residues(struct: Structure, domain_name: str,
chain_order: int, start: int,
finish: int) -> Structure:
"""
Return new structure that contains new model (id=1), new chain (id=1) with residues
from 'start' to 'finish' of specified chain of input structure
:param struct: input structure to slice
:param chain_order: order of chain to extract residues
:param start: start residue
:param finish: finish residues
:param domain_name: new structure name
:return: new structure
"""
new_chain = Chain.Chain(1)
chain = list(struct.get_chains())[chain_order]
for i in range(start, finish + 1):
new_chain.add(chain[i])
model = Model.Model(1)
model.add(new_chain)
domain = Structure.Structure(domain_name)
domain.add(model)
return domain
def initialize_res(residue):
'''Creates a new structure containing a single amino acid. The type and
geometry of the amino acid are determined by the argument, which has to be
either a geometry object or a single-letter amino acid code.
The amino acid will be placed into chain A of model 0.'''
if isinstance(residue, Geo):
geo = residue
else:
geo = geometry(residue)
segID = 1
AA = geo.residue_name
CA_N_length = geo.CA_N_length
CA_C_length = geo.CA_C_length
N_CA_C_angle = geo.N_CA_C_angle
CA_coord = [0., 0., 0.]
C_coord = [CA_C_length, 0, 0]
N_coord = [
CA_N_length * math.cos(N_CA_C_angle * (math.pi / 180.0)),
CA_N_length * math.sin(N_CA_C_angle * (math.pi / 180.0)), 0
]
N = Atom("N", N_coord, 0.0, 1.0, " ", " N", 0, "N")
CA = Atom("CA", CA_coord, 0.0, 1.0, " ", " CA", 0, "C")
C = Atom("C", C_coord, 0.0, 1.0, " ", " C", 0, "C")
##Create Carbonyl atom (to be moved later)
C_O_length = geo.C_O_length
CA_C_O_angle = geo.CA_C_O_angle
N_CA_C_O_diangle = geo.N_CA_C_O_diangle
carbonyl = calculateCoordinates(N, CA, C, C_O_length, CA_C_O_angle,
N_CA_C_O_diangle)
O = Atom("O", carbonyl, 0.0, 1.0, " ", " O", 0, "O")
if (AA == 'G'):
res = makeGly(segID, N, CA, C, O, geo)
elif (AA == 'A'):
res = makeAla(segID, N, CA, C, O, geo)
elif (AA == 'S'):
res = makeSer(segID, N, CA, C, O, geo)
elif (AA == 'C'):
res = makeCys(segID, N, CA, C, O, geo)
elif (AA == 'V'):
res = makeVal(segID, N, CA, C, O, geo)
elif (AA == 'I'):
res = makeIle(segID, N, CA, C, O, geo)
elif (AA == 'L'):
res = makeLeu(segID, N, CA, C, O, geo)
elif (AA == 'T'):
res = makeThr(segID, N, CA, C, O, geo)
elif (AA == 'R'):
res = makeArg(segID, N, CA, C, O, geo)
elif (AA == 'K'):
res = makeLys(segID, N, CA, C, O, geo)
elif (AA == 'D'):
res = makeAsp(segID, N, CA, C, O, geo)
elif (AA == 'E'):
res = makeGlu(segID, N, CA, C, O, geo)
elif (AA == 'N'):
res = makeAsn(segID, N, CA, C, O, geo)
elif (AA == 'Q'):
res = makeGln(segID, N, CA, C, O, geo)
elif (AA == 'M'):
res = makeMet(segID, N, CA, C, O, geo)
elif (AA == 'H'):
res = makeHis(segID, N, CA, C, O, geo)
elif (AA == 'P'):
res = makePro(segID, N, CA, C, O, geo)
elif (AA == 'F'):
res = makePhe(segID, N, CA, C, O, geo)
elif (AA == 'Y'):
res = makeTyr(segID, N, CA, C, O, geo)
elif (AA == 'W'):
res = makeTrp(segID, N, CA, C, O, geo)
else:
res = makeGly(segID, N, CA, C, O, geo)
cha = Chain('A')
cha.add(res)
mod = Model(0)
mod.add(cha)
struc = Structure('X')
struc.add(mod)
return struc
def compare_interactions(interaction1, interaction2, similar_sequences):
"""
This function takes two structures with two chains each one and a dictionary with chains as keys and keys as
values relating them if they have more than a 95% of similarity and returns 1 if the two interactions are
different and 0 if they are the same interaction.
:param interaction1: one of the interactions you want to compare.
:param interaction2: the other interaction you want to compare.
:param similar_sequences: dictionary which relates sequences by similiarity.
:return: returns true if they are the same and false if they ar enot.
"""
homodimer = False # This variable will be true if the chains in the interaction are more than a 95% similar
chain_list1 = []
chain_list2 = []
for chain in interaction1:
chain_id = similar_sequences[chain].get_id(
) # To identify similar chains in the superimposition we name them
# as the main chain of its type
if chain_id in [x.get_id() for x in chain_list1]:
homodimer = True # if the second chain is similar to the first we change homodimer to true and
chain_list1.append(Chain.Chain(chain_id))
res_counter = 0
for residue in chain:
if 'CA' in [
x.get_id() for x in residue.get_atoms()
]: # for every residue in chain that have an alpha carbon
atom = residue['CA'] # storing the alpha carbon
chain_list1[-1].add(
Residue.Residue(
('', res_counter, ''), residue.get_resname(),
residue.get_segid())) # adding the
# residue
chain_list1[-1][('', res_counter, '')].add(
atom.copy()) # adding a copy of the atom to avoid
# modifiying the original ones
res_counter += 1
if 'P' in [
x.get_id() for x in residue.get_atoms()
]: # for every residue in chain that have an alpha carbon
atom = residue['P'] # storing the alpha carbon
chain_list1[-1].add(
Residue.Residue(
('', res_counter, ''), residue.get_resname(),
residue.get_segid())) # adding the
# residue
chain_list1[-1][('', res_counter, '')].add(
atom.copy()) # adding a copy of the atom to avoid
# modifiying the original ones
res_counter += 1
for chain in interaction2: # Doing the same for the structure 2
chain_id = similar_sequences[chain].get_id()
chain_list2.append(Chain.Chain(chain_id))
res_counter = 0
for residue in chain:
if 'CA' in [x.get_id() for x in residue.get_atoms()]:
atom = residue['CA']
chain_list2[-1].add(
Residue.Residue(
('', res_counter, ''), residue.get_resname(),
residue.get_segid()))
chain_list2[-1][('', res_counter, '')].add(atom.copy())
res_counter += 1
if 'P' in [x.get_id() for x in residue.get_atoms()]:
atom = residue['P']
chain_list2[-1].add(
Residue.Residue(
('', res_counter, ''), residue.get_resname(),
residue.get_segid()))
chain_list2[-1][('', res_counter, '')].add(atom.copy())
res_counter += 1
if homodimer: # if the chain is an homodimer we remove different residues from chains in the same interaction
for int in [chain_list1, chain_list2]:
trim_to_superimpose(int[0], int[1])
for chain1 in chain_list1: # Removing different residues betwen similar chains in different interactions
for chain2 in chain_list2:
if chain1.get_id() != chain2.get_id():
continue
trim_to_superimpose(chain1, chain2)
result = str_comparison_superimpose(chain_list1, chain_list2)
return result
if invalid:
continue
num_items = {"O": 0, "T": 0}
for hit in item:
num_items[hit[5]] += 1
for hit in item:
if len(hit[1]) > 1:
structure = PDBParser(QUIET=False).get_structure(
"dimer", path_structure + "dimer.pdb")
new_model = Model(2)
new_chain = Chain("C")
new_model.add(new_chain)
new_chain_mark = Chain("M")
add_markers(new_chain_mark)
new_model.add(new_chain_mark)
structure.add(new_model)
# mutations = {} [sorted IDs] : [cobalt], [residues], num_HIS, clash_info_list, num, type, coords
res_co = Residue((" ", runtime_mark_id, " "), "Co3",
" ")
res_co.add(hit[0])
new_chain.add(res_co) # Add cobalt in its own residue
for res in hit[1]: # Add rotamers
if res not in new_chain:
def initialize_res(residue):
'''Creates a new structure containing a single amino acid. The type and
geometry of the amino acid are determined by the argument, which has to be
either a geometry object or a single-letter amino acid code.
The amino acid will be placed into chain A of model 0.'''
if isinstance( residue, Geo ):
geo = residue
else:
geo=geometry(residue)
segID=1
AA= geo.residue_name
CA_N_length=geo.CA_N_length
CA_C_length=geo.CA_C_length
N_CA_C_angle=geo.N_CA_C_angle
CA_coord= [0.,0.,0.]
C_coord= [CA_C_length,0,0]
N_coord = [CA_N_length*math.cos(N_CA_C_angle*(math.pi/180.0)),CA_N_length*math.sin(N_CA_C_angle*(math.pi/180.0)),0]
N= Atom("N", N_coord, 0.0 , 1.0, " "," N", 0, "N")
CA=Atom("CA", CA_coord, 0.0 , 1.0, " "," CA", 0,"C")
C= Atom("C", C_coord, 0.0, 1.0, " ", " C",0,"C")
##Create Carbonyl atom (to be moved later)
C_O_length=geo.C_O_length
CA_C_O_angle=geo.CA_C_O_angle
N_CA_C_O_diangle=geo.N_CA_C_O_diangle
carbonyl=calculateCoordinates(N, CA, C, C_O_length, CA_C_O_angle, N_CA_C_O_diangle)
O= Atom("O",carbonyl , 0.0 , 1.0, " "," O", 0, "O")
if(AA=='G'):
res=makeGly(segID, N, CA, C, O, geo)
elif(AA=='A'):
res=makeAla(segID, N, CA, C, O, geo)
elif(AA=='S'):
res=makeSer(segID, N, CA, C, O, geo)
elif(AA=='C'):
res=makeCys(segID, N, CA, C, O, geo)
elif(AA=='V'):
res=makeVal(segID, N, CA, C, O, geo)
elif(AA=='I'):
res=makeIle(segID, N, CA, C, O, geo)
elif(AA=='L'):
res=makeLeu(segID, N, CA, C, O, geo)
elif(AA=='T'):
res=makeThr(segID, N, CA, C, O, geo)
elif(AA=='R'):
res=makeArg(segID, N, CA, C, O, geo)
elif(AA=='K'):
res=makeLys(segID, N, CA, C, O, geo)
elif(AA=='D'):
res=makeAsp(segID, N, CA, C, O, geo)
elif(AA=='E'):
res=makeGlu(segID, N, CA, C, O, geo)
elif(AA=='N'):
res=makeAsn(segID, N, CA, C, O, geo)
elif(AA=='Q'):
res=makeGln(segID, N, CA, C, O, geo)
elif(AA=='M'):
res=makeMet(segID, N, CA, C, O, geo)
elif(AA=='H'):
res=makeHis(segID, N, CA, C, O, geo)
elif(AA=='P'):
res=makePro(segID, N, CA, C, O, geo)
elif(AA=='F'):
res=makePhe(segID, N, CA, C, O, geo)
elif(AA=='Y'):
res=makeTyr(segID, N, CA, C, O, geo)
elif(AA=='W'):
res=makeTrp(segID, N, CA, C, O, geo)
else:
res=makeGly(segID, N, CA, C, O, geo)
cha= Chain('A')
cha.add(res)
mod= Model(0)
mod.add(cha)
struc= Structure('X')
struc.add(mod)
return struc
def add_dummies_to_pdb(pdb_structure, chain_a_res_n, chain_b_res_n, target_len, max_angle):
"""
Takes a PDB structure with two associated chains and one dummy chain
Adds another dummy chain identical to the first one
"Draws" a line from chain_a_res_n to chain_b_res_n (basis line)
"Draws" a cone with _max_angle_ aperture and _target_len_ height
Sample two vectors of from the cone
Moves dummy residues so that their CA atoms locate on sampled positions
:param pdb_structure: Bio.PDB.Structure.Structure
:param chain_a_res_n: int, residue number on the first chain
:param chain_b_res_n: int, residue number on the second chain
:param target_len: distance from the CA atoms to dummy CA positions
:param max_angle: aperture of the cone
:return: side-effects
"""
logger = logging.getLogger()
if len([c for c in pdb_structure.get_chains()]) != 3:
logger.error("Wrong PDB structures passed, expected exactly 3 chains.")
sys.exit(1)
chain_a, chain_b, chain_c = pdb_structure.get_chains()
def get_ca_coord(chain_name, chain, res_n):
chain_res_list = [r for r in chain.get_residues()]
if len(chain_res_list) < res_n or res_n <= 0:
logger.error("Wrong number of residue from chain {}: {}".format(chain_name, res_n))
sys.exit(1)
res = chain_res_list[res_n - 1]
if "CA" not in res.child_dict:
logger.error("No CA atom in the residue {} in chain {}.".format(res_n, chain_name))
sys.exit(1)
ca = res.child_dict["CA"]
return ca.get_coord()
chain_a_ca_coord = get_ca_coord("A", chain_a, chain_a_res_n)
chain_b_ca_coord = get_ca_coord("B", chain_b, chain_b_res_n)
diff_vector = chain_a_ca_coord - chain_b_ca_coord
diff_vector_target_len = diff_vector * (target_len / np.linalg.norm(diff_vector))
chain_a_dummy_basic = chain_a_ca_coord + diff_vector_target_len
chain_b_dummy_basic = chain_b_ca_coord - diff_vector_target_len
c_dummy_perpendicular_vector = None
while c_dummy_perpendicular_vector is None:
c_dummy_perpendicular_vector = sample_perpendicular_vector(diff_vector_target_len, max_angle)
c_dummy_ca_coord = chain_a_dummy_basic + c_dummy_perpendicular_vector
d_dummy_perpendicular_vector = None
while d_dummy_perpendicular_vector is None:
d_dummy_perpendicular_vector = sample_perpendicular_vector(diff_vector_target_len, max_angle)
d_dummy_ca_coord = chain_b_dummy_basic + d_dummy_perpendicular_vector
chain_d = Chain.Chain("D")
if len(chain_c.child_list) != 1:
logger.error("Wrong number of children in chain_c: {}".format(len(chain_c.child_list)))
sys.exit(1)
chain_c_child = chain_c.child_list[0]
chain_d.add(deepcopy(chain_c_child))
chain_d_child = chain_d.child_list[0]
model = pdb_structure.child_dict[0]
model.add(chain_d)
chain_c.transform(np.eye(3), c_dummy_ca_coord - chain_c_child.child_dict["CA"].get_coord())
chain_d.transform(np.eye(3), d_dummy_ca_coord - chain_d_child.child_dict["CA"].get_coord())
def compare_interactions(interaction1, interaction2):
structure1 = Structure.Structure('1')
structure2 = Structure.Structure('2')
structure1.add(Model.Model(0))
structure2.add(Model.Model(0))
homodimer = False
for chain in interaction1:
if len(list(structure1[0].get_chains())) == 1 and compare_chains(
chain,
list(structure1[0].get_chains())[0]):
homodimer = True
structure1[0].add(Chain.Chain(chain.get_id()))
res_counter = 0
for residue in chain:
if 'CA' in [x.get_id() for x in residue.get_atoms()]:
atom = residue['CA']
structure1[0][chain.get_id()].add(
Residue.Residue(
('', res_counter, ''), residue.get_resname(),
residue.get_segid()))
structure1[0][chain.get_id()][('', res_counter,
'')].add(atom.copy())
res_counter += 1
for chain in interaction2:
structure2[0].add(Chain.Chain(chain.get_id()))
res_counter = 0
for residue in chain:
if 'CA' in [x.get_id() for x in residue.get_atoms()]:
atom = residue['CA']
structure2[0][chain.get_id()].add(
Residue.Residue(
('', res_counter, ''), residue.get_resname(),
residue.get_segid()))
structure2[0][chain.get_id()][('', res_counter,
'')].add(atom.copy())
res_counter += 1
if homodimer:
for int in [structure1[0], structure2[0]]:
trim_to_superimpose(
list(int.get_chains())[0],
list(int.get_chains())[1])
for chain1 in structure1[0]:
for chain2 in structure2[0]:
if chain1.get_id() != chain2.get_id():
continue
trim_to_superimpose(chain1, chain2)
# print(list(chain1.get_residues())[0])
# print(list(chain2.get_residues())[0])
# print(list(structure1.get_chains()))
# print(list(structure2.get_chains()))
result = str_comparison_superimpose(structure1, structure2)
return result
def assemble_multiscale_visualization(topology_fn, rmf_fn, pdb_dir,
outprefix=None, chimerax=True,
xl_fn=None):
"""
Render multiscale versions of rigid bodies from PDB files + flexible
beads from RMF files w/o mapped crosslinks.
Args:
topology_fn (str): Topolgy file in pipe-separated-value (PSV) format
as required in integrative modeling using IMP. For details on how
to write a topology file, see:
https://integrativemodeling.org/2.13.0/doc/ref/classIMP_1_1pmi_1_1topology_1_1TopologyReader.html
rmf_fn (str): Name of the RMF file.
pdb_dir (str): Directory containing all the PDB files for the rigid
bodies used in modeling.
outprefix (str, optional): Prefix for output files. Defaults to None.
chimerax (bool, optional): If true, a Chimerax script will be written (extension ".cxc"). Defaults to True.
xl_fn (str, optional): A file containing a XL dataset. Defaults to None.
If this dataset is supplied, then it will be mapped on to the overall
structure with satisfied XLs drawn in blue and violated XLs drawn in red.
A XL dataset should be supplied in a comma-separated-value (CSV) format
containing at least the following fields
protein1, residue1, protein2, residue2, sat
where the last field <sat> is a boolean 1 or 0 depending on whether
the particular XL is satisfied (in the ensemble sense) as a result of the
integrative modeling exercise.
"""
# -------------------------------------------
# read the RMF file and extract all particles
# -------------------------------------------
of = RMF.open_rmf_file_read_only(rmf_fn)
rmf_model = IMP.Model()
hier = IMP.rmf.create_hierarchies(of, rmf_model)[0]
IMP.rmf.load_frame(of, 0)
particles = IMP.core.get_leaves(hier)
rmf_ps = {}
for p in particles:
molname = p.get_parent().get_parent().get_parent().get_name().strip()
name = p.get_name().strip()
coord = IMP.core.XYZ(p).get_coordinates()
rmf_ps[(molname, name)] = coord
# --------------------------------------------------------------
# map pdb residues to rmf particles for each rigid body pdb file
# --------------------------------------------------------------
# read the topology file
t = TopologyReader(topology_fn, pdb_dir=pdb_dir)
components = t.get_components()
map_pdb2rmf = {}
rigid_body_models = {}
rigid_body_residues = {}
chain_ids = {} # these are matched to the chimerax rmf plugin
chain_id_count = 0
for c in components:
# ignore unstructured residues
if c.pdb_file == "BEADS": continue
mol = c.molname
pdb_prefix = os.path.basename(c.pdb_file).split(".pdb")[0]
chain_id = c.chain
resrange = c.residue_range
offset = c.pdb_offset
r0 = resrange[0] + offset
r1 = resrange[1] + 1 + offset
if mol not in chain_ids:
chain_ids[mol] = string.ascii_uppercase[chain_id_count]
chain_id_count += 1
if pdb_prefix not in map_pdb2rmf:
map_pdb2rmf[pdb_prefix] = {}
this_rigid_body_model = PDBParser().get_structure("x", c.pdb_file)[0]
this_rigid_body_residues = {(r.full_id[2], r.id[1]): r for r in this_rigid_body_model.get_residues()}
rigid_body_models[pdb_prefix] = this_rigid_body_model
rigid_body_residues[pdb_prefix] = this_rigid_body_residues
for r in range(r0, r1):
key = (chain_id, r)
val = (mol, r)
if key in rigid_body_residues[pdb_prefix]:
map_pdb2rmf[pdb_prefix][key] = val
# --------------------------------
# align all pdb files with the rmf
# --------------------------------
print("\nAligning all rigid body structures...")
align = SVDSuperimposer()
for pdb_prefix, mapper in map_pdb2rmf.items():
pdb_coords = []
pdb_atoms = []
rmf_coords = []
residues = rigid_body_residues[pdb_prefix]
for (chain, pdb_res), (mol, rmf_res) in mapper.items():
r = residues[(chain, pdb_res)]
pdb_coords.append(r["CA"].coord)
pdb_atoms.extend([a for a in r.get_atoms()])
rmf_coords.append(rmf_ps[(mol, str(rmf_res))])
pdb_coords = np.array(pdb_coords)
rmf_coords = np.array(rmf_coords)
align.set(rmf_coords, pdb_coords)
align.run()
rotmat, vec = align.get_rotran()
[a.transform(rotmat, vec) for a in pdb_atoms]
# --------------------------
# assemble the composite pdb
# --------------------------
mols = set(sorted([c.molname for c in components]))
print("\nChain IDs by molecule:")
for k, v in chain_ids.items():
print("molecule %s, chain ID %s" % (k, v))
reslists = {mol: [] for mol in mols}
for pdb_prefix, mapper in map_pdb2rmf.items():
residues = rigid_body_residues[pdb_prefix]
for (chain, pdb_res), (mol, rmf_res) in mapper.items():
r = residues[(chain, pdb_res)] ; resid = rmf_res
new_id = (r.id[0], resid, r.id[2])
new_resname = r.resname
new_segid = r.segid
new_atoms = r.get_atoms()
new_residue = Residue.Residue(id=new_id, resname=new_resname, segid=new_segid)
[new_residue.add(a) for a in new_atoms]
reslists[mol].append(new_residue)
composite_model = Model.Model(0)
for mol, chain_id in chain_ids.items():
this_residues = sorted(reslists[mol], key=lambda r: r.id[1])
this_chain = Chain.Chain(chain_id)
[this_chain.add(r) for r in this_residues]
composite_model.add(this_chain)
# save the composite pdb to file
io = PDBIO()
io.set_structure(composite_model)
if outprefix is None:
outprefix = "centroid_model"
io.save(outprefix + ".pdb")
# -------------------------------------------------------------------
# chimerax rendering (hide most of the rmf except unstructured beads)
# -------------------------------------------------------------------
if not chimerax: exit()
print("\nWriting UCSF Chimerax script...")
s = ""
s += "open %s\n" % (outprefix + ".pdb")
s += "open %s\n" % rmf_fn
s += "hide\n"
s += "show cartoon\n"
s += "color #%d %s\n" % (CHIMERAX_PDB_MODEL_NUM, STRUCT_COLOR)
s += "color #%d %s\n" % (CHIMERAX_RMF_MODEL_NUM, UNSTRUCT_COLOR)
s += "hide #%d\n" % CHIMERAX_RMF_MODEL_NUM
struct_residues = []
for key, val in map_pdb2rmf.items():
struct_residues.extend(list(val.values()))
unstruct_atomspec = {}
for p in rmf_ps:
molname, particle_name = p
rmf_chain_id = chain_ids[molname]
if "bead" in particle_name:
r0, r1 = particle_name.split("_")[0].split("-")
r0 = int(r0) ; r1 = int(r1)
this_atomspec = "#%d/%s:%d-%d" % \
(CHIMERAX_RMF_MODEL_NUM, rmf_chain_id, r0, r1)
for r in range(r0, r1+1):
unstruct_atomspec[(molname, r)] = this_atomspec
else:
if (molname, int(particle_name)) not in struct_residues:
r = int(particle_name)
this_atomspec = "#%d/%s:%d" % \
(CHIMERAX_RMF_MODEL_NUM, rmf_chain_id, r)
unstruct_atomspec[(molname, r)] = this_atomspec
s += "show %s\n" % (" ".join(set(unstruct_atomspec.values())))
# ----------------------------------------------------------
# if crosslink data is supplied, write out a pseudobond file
# ----------------------------------------------------------
if xl_fn is not None:
# parse XL data
df = pd.read_csv(os.path.abspath(xl_fn))
xls = []
for i in range(len(df)):
this_df = df.iloc[i]
p1 = this_df["protein1"] ; r1 = this_df["residue1"]
p2 = this_df["protein2"] ; r2 = this_df["residue2"]
sat = this_df["sat"]
xls.append((p1, r1, p2, r2, sat))
# get lists of struct atomspecs
atomspec = {}
for (mol, particle_name) in rmf_ps:
if "bead" in particle_name: continue
if (mol, int(particle_name)) in unstruct_atomspec: continue
chain_id = chain_ids[mol]
resid = int(particle_name)
atomspec[(mol, resid)] = "#%d/%s:%d@CA" % \
(CHIMERAX_PDB_MODEL_NUM, chain_id, resid)
# now add in all the unstruct atomspecs
atomspec.update(unstruct_atomspec)
# write pseudobond script
s_pb = ""
s_pb += "; radius = %2.2f\n" % XL_RADIUS
s_pb += "; dashes = 0\n"
for xl in xls:
p1, r1, p2, r2, sat = xl
atomspec_1 = atomspec[(p1, r1)]
atomspec_2 = atomspec[(p2, r2)]
if atomspec_1 == atomspec_2:
continue
color = SAT_XL_COLOR if sat else VIOL_XL_COLOR
s_pb += "%s %s %s\n" % (atomspec_1, atomspec_2, color)
s_pb += "\n"
pb_fn = outprefix + "_XLs.pb"
with open(pb_fn, "w") as of:
of.write(s_pb)
s += "open %s\n" % pb_fn
s += "preset 'overall look' publication\n"
chimerax_out_fn = outprefix + ".cxc"
with open(chimerax_out_fn, "w") as of:
of.write(s)
def main():
usage = './align_stems.py [stem_length]'
usage += 'Do diagnostics on the stem model'
parser = OptionParser()
parser.add_option('-i',
'--iterations',
dest='iterations',
default=1,
help="The number of times to repeat the alignment",
type='int')
parser.add_option('-l',
'--length',
dest='length',
default=2,
help="The length of the stem",
type='int')
parser.add_option('-o',
'--output-pdb',
dest='output_pdb',
default=False,
help="Output the structures to pdb files",
action='store_true')
parser.add_option(
'-f',
'--from',
dest='from_file',
default=None,
help='Specify a file to align from. Invalidates the -l option.',
type='str')
parser.add_option(
'-t',
'--to',
dest='to_file',
default=None,
help='Specify a file to align to. Invalidates the -l option.',
type='str')
parser.add_option(
'-m',
'--method',
dest='method',
default='e',
help=
'Specify which method to use for the helix fitting. e = estimate (original, least accurate method), a = align (better, more accurate method), t = template (best, most accurate method)'
)
parser.add_option(
'-a',
'--average-twist',
dest='use_average_method',
default=False,
action='store_true',
help='Use the average of the two twists to align the stems.')
(options, args) = parser.parse_args()
if len(args) < 0:
parser.print_help()
sys.exit(1)
stem_length = options.length
if len(args) == 1:
stem_length = int(args[0])
if options.from_file == None or options.to_file == None:
sss = cbs.get_stem_stats(
os.path.join(cbc.Configuration.base_dir,
'fess/stats/temp.1jj2.stats'))
rmsds = []
for i in range(options.iterations):
if options.from_file != None:
filename = options.from_file
stem_def = stem_def_from_filename(filename)
else:
stem_def = random.choice(sss[stem_length])
filename = '%s_%s.pdb' % (stem_def.pdb_name, "_".join(
map(str, stem_def.define)))
pdb_file = os.path.join(cbc.Configuration.stem_fragment_dir, filename)
# Extract the PDB coordinates of the original chain
with warnings.catch_warnings():
warnings.simplefilter("ignore")
try:
chain = list(bpdb.PDBParser().get_structure(
'temp', pdb_file).get_chains())[0]
chain = cbm.extract_stem_from_chain(chain, stem_def)
except IOError as ie:
cud.pv('ie')
# Convert the chain into a stem model
# This is where the method for fitting a helix is applied
#m = cbm.define_to_stem_model(chain, stem_def.define)
stem = cbm.StemModel(name=stem_def.define)
define = stem_def.define
mids = cgg.get_mids(chain, define, options.method)
stem.mids = tuple([m.get_array() for m in mids])
stem.twists = cgg.get_twists(chain, define)
m = stem
# Create a new chain by aligning the stem from the sampled define
# to the model created from the original stem
new_chain = bpdbc.Chain(' ')
try:
if options.to_file != None:
new_stem_def = stem_def_from_filename(options.to_file)
else:
new_stem_def = random.choice(sss[stem_def.bp_length])
cbm.reconstruct_stem_core(new_stem_def, stem_def.define, new_chain,
dict(), m, options.use_average_method)
except IOError as ie:
cud.pv('ie')
if options.output_pdb:
rtor.output_chain(chain, 'out1.pdb')
rtor.output_chain(new_chain, 'out3.pdb')
unsuperimposed_rmsd = cup.pdb_rmsd(chain,
new_chain,
sidechains=False,
superimpose=False)
superimposed_rmsd = cup.pdb_rmsd(chain,
new_chain,
sidechains=False,
superimpose=True,
apply_sup=True)
rmsds += [[superimposed_rmsd[1], unsuperimposed_rmsd[1]]]
#cud.pv('(superimposed_rmsd, unsuperimposed_rmsd)')
if options.output_pdb:
rtor.output_chain(new_chain, 'out2.pdb')
pp = cvp.PymolPrinter()
(p, n) = m.mids
pp.add_stem_like_core(m.mids, m.twists, stem_def.bp_length + 1, '')
pp.stem_atoms(m.mids, m.twists, stem_def.bp_length + 1)
pp.dump_pymol_file('ss')
print stem_length, superimposed_rmsd[1], unsuperimposed_rmsd[
1], unsuperimposed_rmsd[1] / superimposed_rmsd[1]
