def retrieve_sphere_model(structure): #, score):
"""
each chain is here represented by centre of mass only
"""
sphere_struct = Structure('clustering_model')
my_model = Model(0)
sphere_struct.add(my_model)
#bedzie zmieniona numeracja
chain_mass_centres, index = [], 0
for chain in structure.get_chains():
my_chain = Chain(chain.id)
sphere_struct[0].add(my_chain)
coord = calculate_centre_of_complex(chain)
chain_mass_centres.append(coord)
my_residue = Residue((' ', index, ' '), chain.id, ' ')
coords = array(coord, 'f')
atom = Atom('CA', coords, 0, 0, ' ', 'CA', 1)
my_chain.add(my_residue)
my_residue.add(atom)
index += 1
del structure
return sphere_struct
def renumber(chain, new_id=" "):
"""
Renumber a chain from 1, stripping insertion codes.
:param `Bio.PDB.Chain` chain: structure to sanitise.
:param str new_id: ID of the new chain.
:return: A 2-tuple containing the following:
1. The new :py:class:`Bio.PDB.Chain.Chain` object.
2. A list of tuples containing the old residue ID, as returned by
:py:meth:`Bio.PDB.Chain.Chain.get_id`.
"""
mapping = []
sanitised_chain = Chain(new_id)
for res_index, res in enumerate(chain):
sanitised_res = Residue(
(res.get_id()[0], res_index + 1, ' '),
res.get_resname(),
res.get_segid())
mapping.append(res.get_id())
for atom in res:
sanitised_res.add(atom.copy())
sanitised_chain.add(sanitised_res)
return mapping, sanitised_chain
def getStructFromFasta(self, fname, chainType):
'''
Creates a Bio.PDB.Structure object from a fasta file contained in fname. Atoms are not filled
and thus no coordiantes availables. Implements from Structure to Residue hierarchy.
:param fname: str. path to fasta file
@chainType: str. "l" or "r"
'''
seq = self.parseFasta(
fname, inputNumber="1" if chainType == "l" else
"2") #inpuNumber is used to report which partner fails if error
prefix = self.splitExtendedPrefix(self.getExtendedPrefix(fname))[0]
chainId = chainType.upper()
residues = []
struct = Structure(prefix)
model = Model(0)
struct.add(model)
chain = Chain(chainId)
model.add(chain)
for i, aa in enumerate(seq):
try:
resname = one_to_three(aa)
except KeyError:
resname = "UNK"
res = Residue((' ', i, ' '), resname, prefix)
chain.add(res)
return struct
def renumber_windowed_model(self, structure: Structure, alphafold_mmCIF_dict: Dict) -> Structure:
# Grab the Alphafold dictionary entry that descrives the residue range in the structure
seq_db_align_begin = int(alphafold_mmCIF_dict['_ma_target_ref_db_details.seq_db_align_begin'][0])
seq_db_align_end = int(alphafold_mmCIF_dict['_ma_target_ref_db_details.seq_db_align_end'][0])
# start empty
renumbered_structure = Structure(structure.id)
for model in structure:
renumbered_model = Model(model.id)
for chain in model:
transcript_residue_number = seq_db_align_begin
renumbered_chain = Chain(chain.id)
for residue in chain:
renumbered_residue = residue.copy()
renumbered_residue.id = (' ', transcript_residue_number, ' ')
# The above copy routines fail to copy disorder properly - so just wipe out all notion of disorder
for atom in renumbered_residue:
atom.disordered_flag = 0
renumbered_residue.disordered = 0
renumbered_chain.add(renumbered_residue)
transcript_residue_number += 1
assert transcript_residue_number == seq_db_align_end + 1
renumbered_model.add(renumbered_chain)
renumbered_structure.add(renumbered_model)
return renumbered_structure
def retrieve_ca_model(structure):
"""
chains are represented only by main chain atoms (Calfas or C4')
"""
reduced_struct = Structure('clustering_model')
my_model = Model(0)
reduced_struct.add(my_model)
main_chain_atoms = []
for ch in structure[0]:
my_chain = Chain(ch.id)
reduced_struct[0].add(my_chain)
for resi in ch:
for atom in resi:
#print "----", resi.id, resi.get_segid(), ch.id
if atom.get_name() == "CA" or atom.get_name(
) == "C4'" or atom.get_name() == "C4*":
my_residue = Residue((' ', resi.id[1], ' '),
resi.get_resname(), ' ')
atom = Atom('CA', atom.coord, 0, ' ', ' ', 'CA',
atom.get_serial_number())
my_chain.add(my_residue)
my_residue.add(atom)
main_chain_atoms.append(atom)
return reduced_struct
def create_sphere_representation(self):
"""
each chain is here represented by centre of mass only
"""
new_struct = Structure('sphrere')
my_model = Model(0)
new_struct.add(my_model)
chain_mass_centres, index = [], 1
my_chain = Chain(self.fa_struct.chain)
new_struct[0].add(my_chain)
coord, self.molmass, self.radius = self.calculate_centre_of_complex(
self.fa_struct.struct)
my_residue = Residue((' ', index, ' '), "ALA", ' ')
coords = array(coord, 'f')
atom = Atom('CA', coords, 0, 0, ' ', ' CA', 1)
my_chain.add(my_residue)
my_residue.add(atom)
self.cg_struct = new_struct
name = "dddd" + self.fa_struct.chain
self.save_pdb(new_struct, name)
def normalize_chain(chain: Chain) -> Chain:
new_chain = Chain(chain.id)
for residue in chain:
try:
new_chain.add(normalize_residue(residue))
except UnknownResidueError:
pass
return new_chain
def initialize_res(residue: Union[Geo, str]) -> Structure:
"""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
elif isinstance(residue, str):
geo = geometry(residue)
else:
raise ValueError("Invalid residue argument:", 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 = np.array([0.0, 0.0, 0.0])
C_coord = np.array([CA_C_length, 0, 0])
N_coord = np.array([
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")
# Check if the peptide is capped or not
if geo.residue_name == "ACE":
CA = Atom("CH3", CA_coord, 0.0, 1.0, " ", " CH3", 0, "C")
else:
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")
res = make_res_of_type(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_dummy_structure(self):
"""Adds a dummy atom of zero coordinates to mark a gap in visualisation
software"""
dummy_atom = Atom('DUM', np.zeros(3), 0, 1, ' ', 'DUM', -999)
dummy_residue = Residue((' ', -1 * self.chiral_id, ' '), 'DUM', '?')
dummy_residue.add(dummy_atom)
dummy_chain = Chain('?')
dummy_chain.add(dummy_residue)
self.dummy_structure = dummy_residue
return True
def init_chain(self, chain_id):
"""Initiate a new Chain object with given id.
Arguments:
o chain_id - string
"""
if self.model.has_id(chain_id):
self.chain = self.model[chain_id]
warnings.warn(
"WARNING: Chain %s is discontinuous at line %i." %
(chain_id, self.line_counter), PDBConstructionWarning)
else:
self.chain = Chain(chain_id)
self.model.add(self.chain)
def _align(self):
pp_a = self._pp(self.protein_A, 'A')
# seq_a = pp_a.get_sequence()
pp_b = self._pp(self.protein_B, ' ')
# seq_b = pp_b.get_sequence()
# global_align = pairwise2.align.globalxx(seq_a, seq_b)[0]
# msa = MultipleSeqAlignment([SeqRecord(Seq(global_align[0], alphabet=generic_protein), id='A'),
# SeqRecord(Seq(global_align[1], alphabet=generic_protein), id='B')])
msa = self.alignment
# offset_a = re.search(r'[^-]', str(msa[0].seq)).span()[0]
# offset_b = re.search(r'[^-]', str(msa[1].seq)).span()[0]
plus = 1000
for i in range(len(pp_a)):
pp_a[i].id = (pp_a[i].id[0], plus + i, pp_a[i].id[2])
for i in range(len(pp_b)):
pp_b[i].id = (pp_b[i].id[0], plus + i, pp_b[i].id[2])
new_chain_a = Chain(' ')
for i in pp_a:
# i.id = (i.id[0], i.id[1] - plus, i.id[2])
new_chain_a.add(i)
new_chain_b = Chain(' ')
for i in pp_b:
# i.id = (i.id[0], i.id[1] - plus, i.id[2])
new_chain_b.add(i)
io = PDBIO()
io.set_structure(new_chain_a)
io.save(f'.tmp.protein_a.pdb')
io = PDBIO()
io.set_structure(new_chain_b)
io.save(f'.tmp.protein_b.pdb')
def create_new_chain(self, old_struct):
s = Structure(old_struct.chain)
my_model = Model(0)
s.add(my_model)
my_chain = Chain(old_struct.chain)
my_model.add(my_chain) #what if more chains in one component?
return s
def add(self, residue):
"""Add PdbResidue object to site (in the residues list and dict)"""
residue = residue.copy(include_structure=True)
if type(residue) == PdbResidue:
self.residues.append(residue)
self.residues_dict[residue.full_id] = residue
residue.parent_site = self
if type(residue) == Het:
self.ligands.append(residue)
residue.parent_site = self
if residue.is_polymer:
if residue.chain in self.structure[0]:
for r in residue.structure:
self.structure[0][residue.chain].add(r)
return True
self.structure[0].add(residue.structure)
return True
if residue.structure:
# Initialize structure if empty
if self.structure is None:
self.structure = Structure(self.id)
self.structure.add(Model(0))
chain_id = residue.structure.get_parent().get_id()
if chain_id not in self.structure[0]:
self.structure[0].add(Chain(chain_id))
# Add residue structure to site structure
if residue.structure.get_id() not in self.structure[0][chain_id]:
self.structure[0][chain_id].add(residue.structure)
return True
def slice(cls, obj, selection, name='slice'):
"""Create a new Structure object 'S2' from a slice of the current one, 'S1'. <selection>
defines which descendents 'S1' will be stored in 'S2'."""
from Bio.PDB.Structure import Structure
from Bio.PDB.Model import Model
from Bio.PDB.Chain import Chain
ent = Structure(name) # Biopython structure object
# Loop over selection and determine what model/chain objects we need to create in order to
# store the slice
models = {}
for item in selection:
mid = item[1]
cid = item[2]
if mid not in models:
models[mid] = set() # store chain ids
models[mid].add(cid)
# Create model/chains to store slice
for mid in models:
ent.add(Model(mid))
for cid in models[mid]:
ent[mid].add(Chain(cid))
# Add residues to slice
for item in selection:
mid = item[1]
cid = item[2]
rid = item[3]
ent[mid][cid].add(obj[mid][cid][rid].copy())
return cls(ent, name=name)
def __create_superimposed_pdb(self):
def fill_in_chain(chain, protein_id, rotation_matrix = None):
for index,residue in enumerate(self.proteins[protein_id].get_residues()):
residue.id = (residue.id[0], index, residue.id[2])
chain.add(residue)
merged_model = Model(0)
chain_a = Chain('A')
chain_b = Chain('B')
fill_in_chain(chain_a, 0)
fill_in_chain(chain_b, 1)
merged_model.add(chain_a)
merged_model.add(chain_b)
return merged_model
def from_label_seq_ids(cls, label_seq_ids: Iterable[int],
mapping: BiopythonToMmcifResidueIds.Mapping,
bio_chain: Chain):
return cls(
cls.residues_from_label_seq_ids(label_seq_ids, mapping, bio_chain),
bio_chain.get_parent().get_parent().id,
bio_chain.id,
)
def create_new_chain(self, id):
"""
"""
self.fragment_lattice = Structure(id)
my_model = Model(0)
self.fragment_lattice.add(my_model)
my_chain = Chain(id)
my_model.add(my_chain) #what if more chains in one component?
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= Geo(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= np.array([0.,0.,0.])
C_coord= np.array([CA_C_length,0,0])
N_coord = np.array([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")
res=makeRes(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 get_structure(self, name='RNA chain'):
"""Returns chain as a PDB.Structure object."""
struc = Structure(name)
model = Model(0)
chain = Chain(self.chain_name)
struc.add(model)
struc[0].add(chain)
for resi in self:
struc[0][self.chain_name].add(resi)
return struc
def setUp(self):
self.chain = Chain("A")
residues = [
Residue(0, resname="Trp", segid=0),
Residue(0, resname="His", segid=1),
Residue(0, resname="Ser", segid=2),
Residue(0, resname="Val", segid=3),
Residue(0, resname="His", segid=4),]
for r in residues:
self.chain.add(r)
def add_chain_to_struct(self, chain_id):
"""
adds another model to BIO.pdb structure object
Parameters:
-----------
chain_id : chain name
Returns:
---------
self.struct : Bio.PDB structure with new chain
"""
chain = Chain(chain_id)
self.struct[0].add(chain)
def select_structure(selector, structure):
new_structure = Structure(structure.id)
for model in structure:
if not selector.accept_model(model):
continue
new_model = Model(model.id, model.serial_num)
new_structure.add(new_model)
for chain in model:
if not selector.accept_chain(chain):
continue
new_chain = Chain(chain.id)
new_model.add(new_chain)
for residue in chain:
if not selector.accept_residue(residue):
continue
new_residue = Residue(residue.id, residue.resname,
residue.segid)
new_chain.add(new_residue)
for atom in residue:
if selector.accept_atom(atom):
new_residue.add(atom)
return new_structure
def create_structure(coords, pdb_type, remove_masked):
"""Create the structure.
Args:
coords: 3D coordinates of structure
pdb_type: predict or actual structure
remove_masked: whether to include masked atoms. If false,
the masked atoms have coordinates of [0,0,0].
Returns:
structure
"""
name = protein.id_
structure = Structure(name)
model = Model(0)
chain = Chain('A')
for i, residue in enumerate(protein.primary):
residue = AA_LETTERS[residue]
if int(protein.mask[i]) == 1 or remove_masked == False:
new_residue = Residue((' ', i + 1, ' '), residue, ' ')
j = 3 * i
atom_list = ['N', 'CA', 'CB']
for k, atom in enumerate(atom_list):
new_atom = Atom(name=atom,
coord=coords[j + k, :],
bfactor=0,
occupancy=1,
altloc=' ',
fullname=" {} ".format(atom),
serial_number=0)
new_residue.add(new_atom)
chain.add(new_residue)
model.add(chain)
structure.add(model)
io = PDBIO()
io.set_structure(structure)
io.save(save_dir + name + '_' + pdb_type + '.pdb')
return structure
def createPDBFile(self):
"Create test CIF file with 12 Atoms in icosahedron vertexes"
from Bio.PDB.Structure import Structure
from Bio.PDB.Model import Model
from Bio.PDB.Chain import Chain
from Bio.PDB.Residue import Residue
from Bio.PDB.Atom import Atom
from Bio.PDB.mmcifio import MMCIFIO
import os
CIFFILENAME = "/tmp/out.cif"
# create atom struct with ico simmety (i222r)
icosahedron = Icosahedron(circumscribed_radius=100, orientation='222r')
pentomVectorI222r = icosahedron.getVertices()
# create biopython object
structure = Structure('result') # structure_id
model = Model(1, 1) # model_id,serial_num
structure.add(model)
chain = Chain('A') # chain Id
model.add(chain)
for i, v in enumerate(pentomVectorI222r, 1):
res_id = (' ', i, ' ') # first arg ' ' -> aTOm else heteroatom
res_name = "ALA" #+ str(i) # define name of residue
res_segid = ' '
residue = Residue(res_id, res_name, res_segid)
chain.add(residue)
# ATOM name, coord, bfactor, occupancy, altloc, fullname, serial_number,
# element=None)
atom = Atom('CA', v, 0., 1., " ", " CA ", i, "C")
residue.add(atom)
io = MMCIFIO()
io.set_structure(structure)
# delete file if exists
if os.path.exists(CIFFILENAME):
os.remove(CIFFILENAME)
io.save(CIFFILENAME)
return CIFFILENAME
def init_chain(self, chain_id):
"""Initiate a new Chain object with given id.
Arguments:
o chain_id - string
"""
if self.model.has_id(chain_id):
self.chain=self.model[chain_id]
warnings.warn("WARNING: Chain %s is discontinuous at line %i."
% (chain_id, self.line_counter),
PDBConstructionWarning)
else:
self.chain=Chain(chain_id)
self.model.add(self.chain)
def renumberChain(self,
chainID,
offset=0,
modelID='0',
filename="output.mmcif"):
# get chain object
chain = self.structure[modelID][chainID]
# remove chain from model
self.structure[modelID].detach_child(chainID)
from Bio.PDB.Chain import Chain
# create new chain
newChain = Chain(chainID)
for residue in chain:
# remove residue, otherwise we cannot renumber it
residue.detach_parent()
rId = residue.id
res_id = list(rId)
res_id[1] = res_id[1] + offset
if res_id[1] < 0:
raise ValueError('Residue number cant be <= 0')
residue.id = tuple(res_id)
newChain.add(residue)
self.structure[modelID].add(newChain)
self.write(filename)
def create_new_structure(self, name, chain_id):
"""
creates new Bio.PDB structure object
Parameters:
-----------
name : structure name
chain_id : chain name (e.g. A, B, C)
Returns:
---------
self.struct : Bio.PDB object with model and chain inside
"""
self.struct = Structure(name)
my_model = Model(0)
my_chain = Chain(chain_id)
self.struct.add(my_model)
self.struct[0].add(my_chain)
def __make_structure_from_residues__(self, residues):
"""
Makes a Structure object either from a pdbfile or a list of residues
"""
# KR: this probably can be outsourced to another module.
struct = Structure('s')
model = Model('m')
n_chain = 1
chain = Chain('c%i' % n_chain)
for residue in residues:
if chain.has_id(residue.id):
model.add(chain)
n_chain += 1
chain = Chain('c%i' % n_chain)
chain.add(residue)
model.add(chain)
struct.add(model)
return struct
def polymer(cls,
reslist,
mcsa_id=None,
pdb_id=None,
chain='',
parent_site=None):
"""Alternative constructor for polymers. Takes a residue list and returns
a polymer ligand"""
poly = cls(mcsa_id,
pdb_id,
resname='*P*',
resid=None,
chain=chain,
structure=None,
parent_site=parent_site,
calculate_scores=False)
poly.structure = Chain(chain)
for res in reslist:
if res.get_id() not in poly.structure:
poly.structure.add(res.copy())
poly.similarity, poly.best_match = poly.get_similarity()
poly.centrality = poly.get_centrality()
return poly
class StructureBuilder:
"""Deals with constructing the Structure object.
The StructureBuilder class is used by the PDBParser classes to
translate a file to a Structure object.
"""
def __init__(self):
"""Initialize the class."""
self.line_counter = 0
self.header = {}
def _is_completely_disordered(self, residue):
"""Return 1 if all atoms in the residue have a non blank altloc (PRIVATE)."""
atom_list = residue.get_unpacked_list()
for atom in atom_list:
altloc = atom.get_altloc()
if altloc == " ":
return 0
return 1
# Public methods called by the Parser classes
def set_header(self, header):
"""Set header."""
self.header = header
def set_line_counter(self, line_counter):
"""Tracks line in the PDB file that is being parsed.
Arguments:
- line_counter - int
"""
self.line_counter = line_counter
def init_structure(self, structure_id):
"""Initialize a new Structure object with given id.
Arguments:
- id - string
"""
self.structure = Structure(structure_id)
def init_model(self, model_id, serial_num=None):
"""Create a new Model object with given id.
Arguments:
- id - int
- serial_num - int
"""
self.model = Model(model_id, serial_num)
self.structure.add(self.model)
def init_chain(self, chain_id):
"""Create a new Chain object with given id.
Arguments:
- chain_id - string
"""
if self.model.has_id(chain_id):
self.chain = self.model[chain_id]
warnings.warn(
"WARNING: Chain %s is discontinuous at line %i."
% (chain_id, self.line_counter),
PDBConstructionWarning,
)
else:
self.chain = Chain(chain_id)
self.model.add(self.chain)
def init_seg(self, segid):
"""Flag a change in segid.
Arguments:
- segid - string
"""
self.segid = segid
def init_residue(self, resname, field, resseq, icode):
"""Create a new Residue object.
Arguments:
- resname - string, e.g. "ASN"
- field - hetero flag, "W" for waters, "H" for
hetero residues, otherwise blank.
- resseq - int, sequence identifier
- icode - string, insertion code
"""
if field != " ":
if field == "H":
# The hetero field consists of H_ + the residue name (e.g. H_FUC)
field = "H_" + resname
res_id = (field, resseq, icode)
if field == " ":
if self.chain.has_id(res_id):
# There already is a residue with the id (field, resseq, icode).
# This only makes sense in the case of a point mutation.
warnings.warn(
"WARNING: Residue ('%s', %i, '%s') redefined at line %i."
% (field, resseq, icode, self.line_counter),
PDBConstructionWarning,
)
duplicate_residue = self.chain[res_id]
if duplicate_residue.is_disordered() == 2:
# The residue in the chain is a DisorderedResidue object.
# So just add the last Residue object.
if duplicate_residue.disordered_has_id(resname):
# The residue was already made
self.residue = duplicate_residue
duplicate_residue.disordered_select(resname)
else:
# Make a new residue and add it to the already
# present DisorderedResidue
new_residue = Residue(res_id, resname, self.segid)
duplicate_residue.disordered_add(new_residue)
self.residue = duplicate_residue
return
else:
if resname == duplicate_residue.resname:
warnings.warn(
"WARNING: Residue ('%s', %i, '%s','%s') already defined "
"with the same name at line %i."
% (field, resseq, icode, resname, self.line_counter),
PDBConstructionWarning,
)
self.residue = duplicate_residue
return
# Make a new DisorderedResidue object and put all
# the Residue objects with the id (field, resseq, icode) in it.
# These residues each should have non-blank altlocs for all their atoms.
# If not, the PDB file probably contains an error.
if not self._is_completely_disordered(duplicate_residue):
# if this exception is ignored, a residue will be missing
self.residue = None
raise PDBConstructionException(
"Blank altlocs in duplicate residue %s ('%s', %i, '%s')"
% (resname, field, resseq, icode)
)
self.chain.detach_child(res_id)
new_residue = Residue(res_id, resname, self.segid)
disordered_residue = DisorderedResidue(res_id)
self.chain.add(disordered_residue)
disordered_residue.disordered_add(duplicate_residue)
disordered_residue.disordered_add(new_residue)
self.residue = disordered_residue
return
self.residue = Residue(res_id, resname, self.segid)
self.chain.add(self.residue)
def init_atom(
self,
name,
coord,
b_factor,
occupancy,
altloc,
fullname,
serial_number=None,
element=None,
pqr_charge=None,
radius=None,
is_pqr=False,
):
"""Create a new Atom object.
Arguments:
- name - string, atom name, e.g. CA, spaces should be stripped
- coord - Numeric array (Float0, size 3), atomic coordinates
- b_factor - float, B factor
- occupancy - float
- altloc - string, alternative location specifier
- fullname - string, atom name including spaces, e.g. " CA "
- element - string, upper case, e.g. "HG" for mercury
- pqr_charge - float, atom charge (PQR format)
- radius - float, atom radius (PQR format)
- is_pqr - boolean, flag to specify if a .pqr file is being parsed
"""
residue = self.residue
# if residue is None, an exception was generated during
# the construction of the residue
if residue is None:
return
# First check if this atom is already present in the residue.
# If it is, it might be due to the fact that the two atoms have atom
# names that differ only in spaces (e.g. "CA.." and ".CA.",
# where the dots are spaces). If that is so, use all spaces
# in the atom name of the current atom.
if residue.has_id(name):
duplicate_atom = residue[name]
# atom name with spaces of duplicate atom
duplicate_fullname = duplicate_atom.get_fullname()
if duplicate_fullname != fullname:
# name of current atom now includes spaces
name = fullname
warnings.warn(
"Atom names %r and %r differ only in spaces at line %i."
% (duplicate_fullname, fullname, self.line_counter),
PDBConstructionWarning,
)
if not is_pqr:
self.atom = Atom(
name,
coord,
b_factor,
occupancy,
altloc,
fullname,
serial_number,
element,
)
elif is_pqr:
self.atom = Atom(
name,
coord,
None,
None,
altloc,
fullname,
serial_number,
element,
pqr_charge,
radius,
)
if altloc != " ":
# The atom is disordered
if residue.has_id(name):
# Residue already contains this atom
duplicate_atom = residue[name]
if duplicate_atom.is_disordered() == 2:
duplicate_atom.disordered_add(self.atom)
else:
# This is an error in the PDB file:
# a disordered atom is found with a blank altloc
# Detach the duplicate atom, and put it in a
# DisorderedAtom object together with the current
# atom.
residue.detach_child(name)
disordered_atom = DisorderedAtom(name)
residue.add(disordered_atom)
disordered_atom.disordered_add(self.atom)
disordered_atom.disordered_add(duplicate_atom)
residue.flag_disordered()
warnings.warn(
"WARNING: disordered atom found with blank altloc before "
"line %i.\n" % self.line_counter,
PDBConstructionWarning,
)
else:
# The residue does not contain this disordered atom
# so we create a new one.
disordered_atom = DisorderedAtom(name)
residue.add(disordered_atom)
# Add the real atom to the disordered atom, and the
# disordered atom to the residue
disordered_atom.disordered_add(self.atom)
residue.flag_disordered()
else:
# The atom is not disordered
residue.add(self.atom)
def set_anisou(self, anisou_array):
"""Set anisotropic B factor of current Atom."""
self.atom.set_anisou(anisou_array)
def set_siguij(self, siguij_array):
"""Set standard deviation of anisotropic B factor of current Atom."""
self.atom.set_siguij(siguij_array)
def set_sigatm(self, sigatm_array):
"""Set standard deviation of atom position of current Atom."""
self.atom.set_sigatm(sigatm_array)
def get_structure(self):
"""Return the structure."""
# first sort everything
# self.structure.sort()
# Add the header dict
self.structure.header = self.header
return self.structure
def set_symmetry(self, spacegroup, cell):
"""Set symmetry."""
pass
class StructureBuilder(object):
"""
Deals with contructing the Structure object. The StructureBuilder class is used
by the PDBParser classes to translate a file to a Structure object.
"""
def __init__(self):
self.line_counter=0
self.header={}
def _is_completely_disordered(self, residue):
"Return 1 if all atoms in the residue have a non blank altloc."
atom_list=residue.get_unpacked_list()
for atom in atom_list:
altloc=atom.get_altloc()
if altloc==" ":
return 0
return 1
# Public methods called by the Parser classes
def set_header(self, header):
self.header=header
def set_line_counter(self, line_counter):
"""
The line counter keeps track of the line in the PDB file that
is being parsed.
Arguments:
o line_counter - int
"""
self.line_counter=line_counter
def init_structure(self, structure_id):
"""Initiate a new Structure object with given id.
Arguments:
o id - string
"""
self.structure=Structure(structure_id)
def init_model(self, model_id, serial_num = None):
"""Initiate a new Model object with given id.
Arguments:
o id - int
o serial_num - int
"""
self.model=Model(model_id,serial_num)
self.structure.add(self.model)
def init_chain(self, chain_id):
"""Initiate a new Chain object with given id.
Arguments:
o chain_id - string
"""
if self.model.has_id(chain_id):
self.chain=self.model[chain_id]
warnings.warn("WARNING: Chain %s is discontinuous at line %i."
% (chain_id, self.line_counter),
PDBConstructionWarning)
else:
self.chain=Chain(chain_id)
self.model.add(self.chain)
def init_seg(self, segid):
"""Flag a change in segid.
Arguments:
o segid - string
"""
self.segid=segid
def init_residue(self, resname, field, resseq, icode):
"""
Initiate a new Residue object.
Arguments:
o resname - string, e.g. "ASN"
o field - hetero flag, "W" for waters, "H" for
hetero residues, otherwise blank.
o resseq - int, sequence identifier
o icode - string, insertion code
"""
if field!=" ":
if field=="H":
# The hetero field consists of H_ + the residue name (e.g. H_FUC)
field="H_"+resname
res_id=(field, resseq, icode)
if field==" ":
if self.chain.has_id(res_id):
# There already is a residue with the id (field, resseq, icode).
# This only makes sense in the case of a point mutation.
warnings.warn("WARNING: Residue ('%s', %i, '%s') "
"redefined at line %i."
% (field, resseq, icode, self.line_counter),
PDBConstructionWarning)
duplicate_residue=self.chain[res_id]
if duplicate_residue.is_disordered()==2:
# The residue in the chain is a DisorderedResidue object.
# So just add the last Residue object.
if duplicate_residue.disordered_has_id(resname):
# The residue was already made
self.residue=duplicate_residue
duplicate_residue.disordered_select(resname)
else:
# Make a new residue and add it to the already
# present DisorderedResidue
new_residue=Residue(res_id, resname, self.segid)
duplicate_residue.disordered_add(new_residue)
self.residue=duplicate_residue
return
else:
# Make a new DisorderedResidue object and put all
# the Residue objects with the id (field, resseq, icode) in it.
# These residues each should have non-blank altlocs for all their atoms.
# If not, the PDB file probably contains an error.
if not self._is_completely_disordered(duplicate_residue):
# if this exception is ignored, a residue will be missing
self.residue=None
raise PDBConstructionException(
"Blank altlocs in duplicate residue %s ('%s', %i, '%s')"
% (resname, field, resseq, icode))
self.chain.detach_child(res_id)
new_residue=Residue(res_id, resname, self.segid)
disordered_residue=DisorderedResidue(res_id)
self.chain.add(disordered_residue)
disordered_residue.disordered_add(duplicate_residue)
disordered_residue.disordered_add(new_residue)
self.residue=disordered_residue
return
residue=Residue(res_id, resname, self.segid)
self.chain.add(residue)
self.residue=residue
def init_atom(self, name, coord, b_factor, occupancy, altloc, fullname,
serial_number=None, element=None):
"""
Initiate a new Atom object.
Arguments:
o name - string, atom name, e.g. CA, spaces should be stripped
o coord - Numeric array (Float0, size 3), atomic coordinates
o b_factor - float, B factor
o occupancy - float
o altloc - string, alternative location specifier
o fullname - string, atom name including spaces, e.g. " CA "
o element - string, upper case, e.g. "HG" for mercury
"""
residue=self.residue
# if residue is None, an exception was generated during
# the construction of the residue
if residue is None:
return
# First check if this atom is already present in the residue.
# If it is, it might be due to the fact that the two atoms have atom
# names that differ only in spaces (e.g. "CA.." and ".CA.",
# where the dots are spaces). If that is so, use all spaces
# in the atom name of the current atom.
if residue.has_id(name):
duplicate_atom=residue[name]
# atom name with spaces of duplicate atom
duplicate_fullname=duplicate_atom.get_fullname()
if duplicate_fullname!=fullname:
# name of current atom now includes spaces
name=fullname
warnings.warn("Atom names %r and %r differ "
"only in spaces at line %i."
% (duplicate_fullname, fullname,
self.line_counter),
PDBConstructionWarning)
atom=self.atom=Atom(name, coord, b_factor, occupancy, altloc,
fullname, serial_number, element)
if altloc!=" ":
# The atom is disordered
if residue.has_id(name):
# Residue already contains this atom
duplicate_atom=residue[name]
if duplicate_atom.is_disordered()==2:
duplicate_atom.disordered_add(atom)
else:
# This is an error in the PDB file:
# a disordered atom is found with a blank altloc
# Detach the duplicate atom, and put it in a
# DisorderedAtom object together with the current
# atom.
residue.detach_child(name)
disordered_atom=DisorderedAtom(name)
residue.add(disordered_atom)
disordered_atom.disordered_add(atom)
disordered_atom.disordered_add(duplicate_atom)
residue.flag_disordered()
warnings.warn("WARNING: disordered atom found "
"with blank altloc before line %i.\n"
% self.line_counter,
PDBConstructionWarning)
else:
# The residue does not contain this disordered atom
# so we create a new one.
disordered_atom=DisorderedAtom(name)
residue.add(disordered_atom)
# Add the real atom to the disordered atom, and the
# disordered atom to the residue
disordered_atom.disordered_add(atom)
residue.flag_disordered()
else:
# The atom is not disordered
residue.add(atom)
def set_anisou(self, anisou_array):
"Set anisotropic B factor of current Atom."
self.atom.set_anisou(anisou_array)
def set_siguij(self, siguij_array):
"Set standard deviation of anisotropic B factor of current Atom."
self.atom.set_siguij(siguij_array)
def set_sigatm(self, sigatm_array):
"Set standard deviation of atom position of current Atom."
self.atom.set_sigatm(sigatm_array)
def get_structure(self):
"Return the structure."
# first sort everything
# self.structure.sort()
# Add the header dict
self.structure.header=self.header
return self.structure
def set_symmetry(self, spacegroup, cell):
pass
class TestPdbalign(unittest.TestCase):
# Need to reduce gap penalty to make test alignments work
aligner = Aligner(BLOSUM62.load(), do_codon=False,
open_insertion=-1, open_deletion=-1)
def setUp(self):
self.chain = Chain("A")
residues = [
Residue(0, resname="Trp", segid=0),
Residue(0, resname="His", segid=1),
Residue(0, resname="Ser", segid=2),
Residue(0, resname="Val", segid=3),
Residue(0, resname="His", segid=4),]
for r in residues:
self.chain.add(r)
def test_align_and_index(self):
problems = (
(Seq("AHSVH"), Seq("AHVH"), [0, 1, -1, 2, 3]),
(Seq("AHVH"), Seq("AHSVH"), [0, 1, 3, 4]),
(Seq("AHSVH"), Seq("AHSVH"), [0, 1, 2, 3, 4]),
(Seq("-HSVH"), Seq("AHSVH"), [-1, 1, 2, 3, 4]),
(Seq("A-SVH"), Seq("AHSVH"), [0, -1, 2, 3, 4]),
(Seq("AH-VH"), Seq("AHSVH"), [0, 1, -1, 3, 4]),
(Seq("AHS-H"), Seq("AHSVH"), [0, 1, 2, -1, 4]),
(Seq("AHSV-"), Seq("AHSVH"), [0, 1, 2, 3, -1]),
(Seq("AHSVHCCCCCCFPVW"), Seq("AHSVHFPVW"),
[0, 1, 2, 3, 4, -1, -1, -1, -1, -1, -1, 5, 6, 7, 8]),
)
for s, p, e in problems:
result = align_and_index(s, p, missing=-1, aligner=self.aligner)
self.assertEqual(e, result)
def test_align_chains_msa(self):
sequences = [Seq("AHSVH"),
Seq("AH-VH"),
Seq("A-SVH")]
indices = align_chains_msa(sequences, [self.chain], aligner=self.aligner)
expected = np.array([[0, 1, 2, 3, 4]])
self.assertTrue(np.all(indices == expected))
def test_align_chains_msa_no_consensus(self):
sequences = [Seq("AHSV"),
Seq("AHSH")]
indices = align_chains_msa(sequences, [self.chain], aligner=self.aligner)
expected = np.array([[0, 1, 2, -1]])
self.assertTrue(np.all(indices == expected))
def test_align_chains_msa_leading_gaps(self):
sequences = [Seq("FFWHSVH"),
Seq("IIWH-VH"),
Seq("WWW-SVH")]
indices = align_chains_msa(sequences, [self.chain], aligner=self.aligner)
expected = np.array([[-1, -1, 0, 1, 2, 3, 4]])
self.assertTrue(np.all(indices == expected))
def test_compute_distance_matrix(self):
c1 = np.array([[0, 0],
[np.nan, np.nan],
[1, 1],
[1, 0]])
c2 = c1.copy()
c1[:, 0] += 1.5
c1[:, 1] += 1
coords = np.hstack([c1, c2]).reshape((4, 2, 2))
expected = np.array([[0, 5, 0.5, 1],
[5, 0, 5, np.inf],
[0.5, 5, 0, 1],
[1, np.inf, 1, 0]])
result = compute_distance_matrix(coords, default_dist=5)
self.assertTrue(np.all(expected == result))
def test_consensus(self):
flag = -1
problems = (((0, 0, 1, 1), flag),
((0, 0, 0, 1), 0),
((0, 0, 0, 0), 0),
(iter([]), flag),
((), flag))
for it, exp in problems:
result = consensus(it, flag=-1)
self.assertEqual(exp, result)
ind = 0
for line in open(filename).readlines():
if not line.startswith('#'):
bfactors[ind] = array((line.split())[column])
ind = ind+1
return bfactors
#--------------------------------------------------------------------
points = ReadXYZ ( args['src'], args['scale'])
if ( args['bfactor'] is not None):
print "read bfactor file column %d" % args['column']
bfactors = ReadBfactor(args['bfactor'],args['column'])
else:
bfactors = zeros(len(points))
model = Model(1)
chain = Chain("A")
structure = Structure("ref")
num_count = 0
for i in range(0,shape(points)[0]):
num_count = num_count +1
res_id = (' ',num_count,' ')
residue = Residue(res_id,'ALA',' ')
cur_coord = tuple(points[i])
bfactor = bfactors[i]
atom = Atom('CA',cur_coord,bfactor,0,' ','CA',num_count,'C')
residue.add(atom)
chain.add(residue)
model.add(chain)
structure.add(model)
def compare_chains(chain1: Chain, chain2: Chain,
c1_residue_mapping: BiopythonToMmcifResidueIds.Mapping,
c2_residue_mapping: BiopythonToMmcifResidueIds.Mapping,
c1_seq: Dict[int, str], c2_seq: Dict[int, str], # in 3-letter codes
comparators__residues_param: List[Analyzer],
comparators__residue_ids_param: List[Analyzer],
comparators__domains__residues_param: List[Analyzer],
comparators__domains__residue_ids_param: List[Analyzer],
comparators__2domains__residues_param: List[Analyzer],
serializer_or_analysis_handler: AnalysisHandler,
domains_info: list,
) -> None:
""" Runs comparisons between two chains. E.g. one ligand-free (apo) and another ligand-bound (holo).
:param chain1: A Bio.PDB Chain, obtained as a part of BioPython Structure object as usual
:param chain2: A corresponding chain (same sequence), typically from a different PDB structure. See chain1.
:param c1_residue_mapping:
:param apo_poly_seqs:
"""
s1_pdb_code = chain1.get_parent().get_parent().id
s2_pdb_code = chain2.get_parent().get_parent().id
logging.info(f'running analyses for ({s1_pdb_code}, {s2_pdb_code}) pair...')
#
# with warnings.catch_warnings():
# warnings.simplefilter("ignore")
# pp1 = chain_to_polypeptide(chain1)
# pp2 = chain_to_polypeptide(chain2)
# c1_seq, c2_seq todo, is the order in atom_site loop guaranteed? If not, I should sort the dict by label_seq_id
# also todo, is label_seq_id sequential, that is one-by-one always +1?
# todo assert entity_poly_seq have no gaps (always +1), they say they're sequential, I think they mean exactly this
# crop polypeptides to longest common substring
c1_common_seq, c2_common_seq = get_longest_common_polypeptide(c1_seq, c2_seq)
c1_label_seq_ids = list(c1_common_seq.keys())
c2_label_seq_ids = list(c2_common_seq.keys())
return
label_seq_id_offset = c2_label_seq_ids[0] - c1_label_seq_ids[0]
# up to this point, we have residue ids of the protein sequence in the experiment. This also includes unobserved
# residues, but those we will exclude from our analysis as their positions weren't determined
c1_residues, c1_label_seq_ids, c2_residues, c2_label_seq_ids = get_observed_residues(
chain1,
c1_label_seq_ids,
c1_residue_mapping,
chain2,
c2_label_seq_ids,
c2_residue_mapping,
)
c1_residues = ChainResidues(c1_residues, s1_pdb_code, chain1.id)
c2_residues = ChainResidues(c2_residues, s2_pdb_code, chain2.id)
# todo trochu nesikovny
c1_residue_ids = ChainResidueData[ResidueId]([ResidueId(label_seq_id, chain1.id) for label_seq_id in
c1_label_seq_ids], s1_pdb_code, chain1.id)
c2_residue_ids = ChainResidueData[ResidueId]([ResidueId(label_seq_id, chain2.id) for label_seq_id in
c2_label_seq_ids], s2_pdb_code, chain2.id)
# [done] tady nahradit pp pomocí apo_seq nějak
# [done] v analyzerech (APIs) nahradit author_seq_id
# todo tady matchovaní domén pomocí tohodle - zas mohu pouzit Sequence Matcher
# - ale spany, je to složitější -> zatím přeindexovat apo nebo holo do druhý...
for a in comparators__residues_param:
# this fn (run_analyses_for_isoform_group) does not know anything about serialization?
# But it will know how nested it is (domain->structure) and can pass full identifiers of structures/domains
serializer_or_analysis_handler.handle('chain2chain', a, a(c1_residues, c2_residues), c1_residues,
c2_residues) # in future maybe pass apo and holo. Will serialize itself. And output the object in rdf for example?
# because what I would like is to output the analysis with objects identifiers, and then output the objects, what they contain (e.g. domain size?)
for c in comparators__residue_ids_param:
serializer_or_analysis_handler.handle('chain2chain', c, c(c1_residue_ids, c2_residue_ids), c1_residue_ids,
c2_residue_ids)
# domain-level analyses
# get domains (set of auth_seq_id), sort them by domain id and hope they will correspond to each other
# or could map corresponding domains by choosing the ones that have the most overlap?
try:
c1_domains = sorted(filter(lambda d: d.chain_id == chain1.id, get_domains(s1_pdb_code)), key=lambda d: d.domain_id)
c2_domains = sorted(filter(lambda d: d.chain_id == chain2.id, get_domains(s2_pdb_code)), key=lambda d: d.domain_id)
# todo zaznamenat total počet domén (pro obě struktury), zapsat do jinýho jsonu třeba
for pdb_code, domains in ((s1_pdb_code, c1_domains), (s2_pdb_code, c2_domains)):
for d in domains:
domains_info.append(
{'type': 'full_domain',
'full_id': (pdb_code, d.chain_id, d.domain_id),
'pdb_code': pdb_code,
'chain_id': d.chain_id,
'domain_id': d.domain_id,
'spans': d.get_spans(),})
# for d in c2_domains:
# domains_info.append(
# {'type': 'total_domains_found', 'result': len(c2_domains), 'pdb_code': s2_pdb_code, 'chain_id': chain2.id})
# todo spany domén, hlavně
except APIException as e:
if e.__cause__ and '404' in str(e.__cause__):
logging.warning(f'{s1_pdb_code} {s2_pdb_code} no domains found, skip the domain-level analysis')
return # no domains found, skip the domain-level analysis
raise
# assert len(c1_domains) == len(c2_domains) # not always true, as expected, but now OK
# SequenceMatcher on domain resiudes
c1_domains__residues = []
c2_domains__residues = []
for c1_d in c1_domains: # or c2_domains:
# first remap first domain to second (or in future use longest common substrings, but not trivial since domains can be composed of multiple segments)
# offset nemusí být všude stejný
c1_domain_mapped_to_c2 = DomainResidueMapping.from_domain_on_another_chain(c1_d, chain2.id, label_seq_id_offset)
# todo proc chain.get_parent?? Asi abych chain nemusel specifikovat (ale ted pracuju jenom s nima..)
c1_d_residues = DomainResidues.from_domain(c1_d, chain1.get_parent(), c1_residue_mapping,
lambda id: id not in c1_label_seq_ids)
c2_d_residues = DomainResidues.from_domain(c1_domain_mapped_to_c2, chain2.get_parent(), c2_residue_mapping,
lambda id: id not in c2_label_seq_ids)
if not c1_d_residues or not c2_d_residues:
# the domain is not within the processed LCS of both chains (empty intersection with chain residues)
logging.warning(f'domain {c1_d.domain_id} is not within the processed LCS of both chains (empty '
f'intersection with '
f'chain residues)')
continue
c1_domains__residues.append(DomainResidues(c1_d_residues.data, c1_d_residues.structure_id, c1_d_residues.chain_id, c1_d_residues.domain_id))
c2_domains__residues.append(DomainResidues(c2_d_residues.data, c2_d_residues.structure_id, c2_d_residues.chain_id, c2_d_residues.domain_id))
for residue_mapping, domains in ((c1_residue_mapping, c1_domains__residues),
(c2_residue_mapping, c2_domains__residues)):
for d in domains:
domains_info.append(
{'type': 'analyzed_domain',
'full_id': d.get_full_id(),
'pdb_code': d.structure_id,
'chain_id': d.chain_id,
'domain_id': d.domain_id,
'spans': d.get_spans(residue_mapping),
'spans_auth_seq_id': d.get_spans(residue_mapping, auth_seq_id=True),
})
#
# # todo zaznamenat počet domén jdoucích do analýz
# domains_info.append({'type': 'analyzed_domain_count', 'result': len(c1_domains__residues), 'pdb_code': s1_pdb_code, 'chain_id': chain1.id})
# domains_info.append({'type': 'analyzed_domain_count', 'result': len(c2_domains__residues), 'pdb_code': s2_pdb_code, 'chain_id': chain2.id})
# todo to tam taky neni v argumentech, ale harcoded.., to je ten muj fix...
# todo tohle totiž neni párový porovnání.., ale 'jednotkový'
# - stejně jako get domains, get_ss (nikoliv compare ss), vlastne i sequence atp
# - cachovat surface area teda nedava smysl, nacte se proste z predvypocitanyho, jako normalne
# - nebo, proste jenom tyhle structure-level veci ma smysl "cachovat" resp nepocitat tady, pro kazdej par, ale
# - nacitat z filu/unpicklovat - to asi ne, mít serialize/deserialize (stejne chci to mit jako citelny vystup). 4
# - A pak to klidně všechno pro rychlost deserializovat do pameti...
# no, tak to abych se těšil zas na json/pandas-merge hell.. Vsude merge.. Vsude dupe cols/delat index (ten pak ale nekdy zas potrebujes v cols...)
for chain_domains in (c1_domains__residues, c2_domains__residues):
for d1, d2 in itertools.combinations(chain_domains, 2):
serializer_or_analysis_handler.handle('2DA', get_interdomain_surface, get_interdomain_surface(d1, d2),
d1, d2)
for d_chain1, d_chain2 in zip(c1_domains__residues, c2_domains__residues):
for a in comparators__domains__residues_param:
serializer_or_analysis_handler.handle('domain2domain', a, a(d_chain1, d_chain2), d_chain1, d_chain2)
# todo vyres ty divny idcka
for d_chain1, d_chain2 in zip(c1_domains__residues, c2_domains__residues):
# Convert DomainResidues to DomainResidueData[ResidueId]
# asi zas přes mapping... lepší by to bylo, kdyby byl implicitně schovaný třeba na to biopython residue (
# jinak by to nešlo moc ani, leda mit CustomResidue s fieldama bioresidue a label_seq_id, to je ale celkem
# naprd, nebo ne? Nefungovalo by to s chainem, ale to stejně nikde nepoužívám...
d_chain1 = DomainResidueData[ResidueId]([ResidueId.from_bio_residue(r, c1_residue_mapping) for r in d_chain1],
d_chain1.structure_id, d_chain1.chain_id, d_chain1.domain_id)
d_chain2 = DomainResidueData[ResidueId]([ResidueId.from_bio_residue(r, c2_residue_mapping) for r in d_chain2],
d_chain2.structure_id, d_chain2.chain_id, d_chain2.domain_id)
for a in comparators__domains__residue_ids_param:
serializer_or_analysis_handler.handle('domain2domain', a, a(d_chain1, d_chain2), d_chain1, d_chain2)
# two-domain arrangements to two-domain arrangements
for (d1_chain1, d1_chain2), (d2_chain1, d2_chain2) in itertools.combinations(zip(c1_domains__residues, c2_domains__residues), 2):
# (in paper considered if of both apo and holo interdomain iface >= 200 A^2
# if get_interdomain_surface(d1_chain1, d2_chain1) < 200 or get_interdomain_surface(d1_chain2, d2_chain2) < 200:
# continue
for a in comparators__2domains__residues_param:
serializer_or_analysis_handler.handle('chain2DA2chain2DA', a, a(d1_chain1, d2_chain1, d1_chain2,
d2_chain2),
d1_chain1,
d2_chain1, d1_chain2, d2_chain2)
d1d2_chain1 = d1_chain1 + d2_chain1
d1d2_chain2 = d1_chain2 + d2_chain2
serializer_or_analysis_handler.handle('chain2DA2chain2DA', get_rmsd, get_rmsd(d1d2_chain1, d1d2_chain2),
d1d2_chain1,
d1d2_chain2) # todo hardcoded analysis
def run(self, struct: Model, chain: Chain):
def has_at_least_n_non_hydrogen_atoms(ligand, n):
non_hydrogen_atoms = 0
for atom in ligand.get_atoms():
assert atom.element is not None
if atom.element != 'H':
non_hydrogen_atoms += 1
if non_hydrogen_atoms >= n:
return True # todo nakonec můžu asi sumovat všechny, stejně budu chtít konfigurovatelny output, aby mi dal počet atomů ligandu, nebo budu dělat statistiky, kolik atomu má průměrný ligand atp.
return False
# ligand has >= 6 non-hydrogen atoms
ligands = list(
filter(lambda lig: has_at_least_n_non_hydrogen_atoms(lig, 6),
get_all_ligands(struct)))
# ligand is within RADIUS in contact with MIN_RESIDUES_WITHIN_LIGAND residues
# (in original paper they used a program LPC, for ensuring specific interaction of ligand with at least 6 residue, this is a "shortcut",
# a temporary condition (simple))
chain_atoms = list(chain.get_atoms())
ns = NeighborSearch(chain_atoms)
RADIUS = 4.5
MIN_RESIDUES_WITHIN_LIGAND = 6
# todo calculate average number of protein heavy atoms in 4.5 Å within ligand atom (paper says 6)
acceptable_ligands = []
for ligand in ligands:
residues_in_contact_with_ligand = set(
) # including the ligand itself (in biopython, non-peptide ligand is
# in the same chain usually, but in a different residue)
ligand_residues = set() # residues that compose the ligand
for ligand_atom in ligand.get_atoms(
): # ligand can be a chain or a residue
ligand_residues.add(ligand_atom.get_parent())
chain_atoms_in_contact = ns.search(ligand_atom.get_coord(),
RADIUS)
for atom in chain_atoms_in_contact:
# exclude hydrogen atoms (as in the paper)
if atom.element == 'H':
continue
residues_in_contact_with_ligand.add(atom.get_parent())
# exclude the ligand itself from the set of contact residues
residues_in_contact_with_ligand -= ligand_residues
if len(residues_in_contact_with_ligand
) >= MIN_RESIDUES_WITHIN_LIGAND:
acceptable_ligands.append(ligand)
return len(acceptable_ligands) > 0
points = zeros(shape=(num_lines,3))
ind = 0
for line in open(filename).readlines():
points[ind] = array((line.split()[0:3]))
points[ind] = points[ind] * scale
ind = ind+1
return points
#--------------------------------------------------------------------
ref_ptsfilename = "K562.pts"
refid = "ref"
structure = Structure(refid)
model_ref = Model(1)
chain_ref = Chain("A")
points_ref = ReadXYZ(ref_ptsfilename,scale)
num_count = 0
for i in range(0,shape(points_ref[IndexList])[0]):
num_count = num_count +1
res_id = (' ',num_count,' ')
residue = Residue(res_id,'ALA',' ')
cur_coord = tuple(points_ref[IndexList[i]])
atom = Atom('CA',cur_coord,0,0,' ',num_count,num_count,'C')
residue.add(atom)
chain_ref.add(residue)
model_ref.add(chain_ref)
structure.add(model_ref)
#--------------------------------------------------------------------
