def check_loop_rmsd(self, query_pose, subject_pose):
"""
If the subject and query loops are the same size, take CA RMSD
"""
if self.query_loop_size == self.subject_loop_size:
self.subject_loop_matches_query_length = True
# Get pose-numbered residues for loop termini
qn = self.N_splice_res.query_pose_number
sn = self.N_splice_res.subject_pose_number
qc = self.C_splice_res.query_pose_number
sc = self.C_splice_res.subject_pose_number
# Make residue selectors for loops
stemp = '{}-{}'
query_selector = ResidueIndexSelector(stemp.format(qn, qc))
subject_selector = ResidueIndexSelector(stemp.format(sn, sc))
# Calculate RMSD
self.rmsd = align_protein_sections(query_pose, query_selector,
subject_pose, subject_selector)
else:
self.subject_loop_matches_query_length = False
return
def process(pr, p, args, out):
selector = ResidueIndexSelector()
for residue in args.residues:
selector.append_index(residue)
hits = pr.search(p, args.bidentates + args.networks, selector)
prev = -1
buffer = []
limit = args.n_cutoff if args.reduced_output else 0
for (hash, pose) in hits:
if (hash != prev):
for n, (_, minimized) in enumerate(filter_clash_minimize(p, buffer, clash_cutoff=args.clash_cutoff, limit=limit)):
print("Match found! Pair hash: %d, match number: %d" % (prev, n))
minimized.dump_pdb(path.join(out, "result_pair%d_num%d.pdb" % (prev, n)))
prev = hash
buffer = []
else:
buffer.append((hash, pose))
if (buffer):
for n, (_, minimized) in enumerate(filter_clash_minimize(p, buffer, clash_cutoff=args.clash_cutoff, limit=limit)):
print("Match found! Pair hash: %d, match number: %d" % (prev, n))
minimized.dump_pdb(path.join(out, "result_pair%d_num%d.pdb" % (prev, n)))
buffer = []
def measure_affinity(pose,
target_residue,
interacting_residues,
score_function="auto"):
"""
"""
assert type(target_residue) == int, \
"'target_residue' should be an int with the residue ID"
if score_function == "auto":
score_function = get_fa_scorefxn()
if type(interacting_residues) == int:
interacting_residues = [interacting_residues]
hbonds_number, hbonds_energy = measure_hbonds_number(
pose, target_residue, interacting_residues)
target = ResidueIndexSelector(str(target_residue))
interface = ResidueIndexSelector(ID2Rank(pose, interacting_residues))
try:
interaction_energy_metric = InteractionEnergyMetric(target, interface)
except:
from pyrosetta.rosetta.core.simple_metrics.metrics import \
InteractionEnergyMetric
interaction_energy_metric = InteractionEnergyMetric(target, interface)
results = {}
results["total_energy"] = score_function(pose)
results["hbonds_number"] = hbonds_number
results["hbonds_energy"] = hbonds_energy
results["inter_energy"] = interaction_energy_metric.calculate(pose)
return results
def make_task_factory(residue_selectors, confine_design=None):
"""
Makes a TaskFactory with operations that leave the mutable residues
designable, restricts the nearby residues to repacking, and prevents
repacking of other residues.
Also includes the ability to take in a target residue mutatations text
file in the form:
res_number allowed_AAs
for example:
138 KR
All designable residues not listed in the file are restricted to repacking
and the listed residues are limited in their design options to those AAs
listed.
"""
design_set = residue_selectors['mutable']
repack_set = residue_selectors['packable']
other_set = residue_selectors['other']
prevent = PreventRepackingRLT() # No repack, no design
repack = RestrictToRepackingRLT() # No design
tf = TaskFactory()
tf.push_back(OperateOnResidueSubset(prevent, other_set))
tf.push_back(OperateOnResidueSubset(repack, repack_set))
# Everything else left designable by default
# Restricting design further
if confine_design:
trms = readfile(confine_design)
# Converting lines to a dict
limited_set = \
{int(line.split()[0]):line.split()[1] for line in trms}
# Converting residues in the dict to a selector
res_concatenated = str(limited_set.keys()).strip('[]').replace(' ','')
small_des_set = ResidueIndexSelector(res_concatenated)
now_only_repack = NotResidueSelector(small_des_set)
# Making residue selection excluding residues in the file and
# restricting them to repacking
no_longer_designable = AndResidueSelector()
no_longer_designable.add_residue_selector(design_set)
no_longer_designable.add_residue_selector(now_only_repack)
tf.push_back(OperateOnResidueSubset(repack, no_longer_designable))
# Limiting design on residues in the file
for res, AAs in limited_set.items():
designable_res = ResidueIndexSelector(str(res))
restrict_AAs = RestrictAbsentCanonicalAASRLT()
restrict_AAs.aas_to_keep(AAs)
tf.push_back(OperateOnResidueSubset(restrict_AAs, designable_res))
return tf
def select_residues(cat_res,
peptide_subset,
design_protease=True,
design_peptide=False):
"""
Makes residue selectors for protease sections. Requires manual input for
which residues are catalytic and whether only part of the peptide should be
selected. Options for whether the peptide is designable (false by default)
and whether the protease is designable (true by default). Assumes that the
protease is chain A and the peptide is chain B.
"""
residue_selectors = {}
# Protease residues. Protease assumed to be chain A
protease = ChainSelector("A")
residue_selectors['protease'] = protease
# Peptide residues. Peptide assumed to be chain B, unless range specified
if peptide_subset:
peptide = ResidueIndexSelector(peptide_subset)
else:
peptide = ChainSelector("B")
residue_selectors['peptide'] = peptide
# Catalytic residues. ResidueIndexSelector needs a string, not a list.
if cat_res:
cats_as_str = ','.join([str(i) for i in cat_res])
catalytic = ResidueIndexSelector(cats_as_str)
else:
# If no catalytic residues are given, return a null selector
catalytic = selector_intersection(protease, peptide) # Empty set
residue_selectors['catalytic'] = catalytic
# Designable residues. May include protease and peptide, just one, or none
if design_protease:
mutable = mutable_residues_selector(protease, peptide, catalytic,
design_peptide)
elif design_peptide:
mutable = peptide
else: # Neither protease not peptide designable
mutable = selector_intersection(protease, peptide) # Empty set
residue_selectors['mutable'] = mutable
# Packable residues. Centered around the peptide and designable set
packable = packable_residues_selector(peptide, mutable, catalytic)
residue_selectors['packable'] = packable
# Immobile residues. Catalytic residues and everything that isn't mutable
# or packable
immobile = NotResidueSelector(selector_union(mutable, packable))
residue_selectors['immobile'] = immobile
return residue_selectors
def create_design_task_factory(designable_res_indexes: list, active_site_positions: list = list(), \
neighborhood: float = 0, cystine: bool = True, noncanonical_amino_acids: list = list()):
task_factory = TaskFactory()
if len(noncanonical_amino_acids) > 0:
ncaa_palette = CustomBaseTypePackerPalette()
for ncaa in noncanonical_amino_acids:
ncaa_palette.add_type(ncaa)
task_factory.set_packer_palette(ncaa_palette)
task_factory.push_back(IncludeCurrent())
# Mutatable
designable_selector = ResidueIndexSelector(','.join(
str(designable_res_index)
for designable_res_index in designable_res_indexes))
if not cystine:
restriction = RestrictAbsentCanonicalAASRLT()
restriction.aas_to_keep('AGILPVFWYDERHKSTMNQ') # AGILPVFWYDERHKSTCMNQ
task_factory.push_back(
OperateOnResidueSubset(restriction, designable_selector))
# Repack
repack = RestrictToRepackingRLT()
prevent = PreventRepackingRLT()
if neighborhood > 0:
if len(active_site_positions) > 0:
enzdes_cst_selector = ResidueIndexSelector(','.join(
str(enzdes_cst_index)
for enzdes_cst_index in active_site_positions))
designable_repacking_selector = NeighborhoodResidueSelector()
designable_repacking_selector.set_focus_selector(
OrResidueSelector(designable_selector, enzdes_cst_selector))
designable_repacking_selector.set_distance(neighborhood)
designable_repacking_selector.set_include_focus_in_subset(True)
repacking_selector = AndResidueSelector(
designable_repacking_selector,
NotResidueSelector(designable_selector))
task_factory.push_back(
OperateOnResidueSubset(repack, repacking_selector))
task_factory.push_back(
OperateOnResidueSubset(prevent, designable_repacking_selector,
True))
else:
repacking_selector = NeighborhoodResidueSelector()
repacking_selector.set_focus_selector(designable_selector)
repacking_selector.set_distance(neighborhood)
repacking_selector.set_include_focus_in_subset(False)
task_factory.push_back(
OperateOnResidueSubset(repack, repacking_selector))
task_factory.push_back(OperateOnResidueSubset(prevent, \
OrResidueSelector(designable_selector, repacking_selector), True))
else:
task_factory.push_back(
OperateOnResidueSubset(repack, designable_selector, True))
return task_factory
def get_mobile_range(chain_no, central_res=None, range_from_tug=None):
"""
Returns a selection of residues in given chain. The list can either include
the full chain, or a subset within a given range from a given central
residue. Still includes stom hard-coding for tau in the case of the subset.
"""
# Full chain
if not all([central_res, range_from_tug]):
chain_letter = chr(64 + chain_no) # 'A' is chr 65, 'B' is 66, etc.
chain_selection = ChainSelector(chain_letter)
return chain_selection
# Subset of chain
elif all([central_res, range_from_tug]):
# Ignoring upstream residues of central res is near N-term
if central_res >= 306 + range_from_tug:
run_start = range_from_tug
else:
run_start = central_res - 306
# Ignoring downstream residues if central res is near C-term
if central_res <= 378 - range_from_tug:
run_end = range_from_tug
else:
run_end = 378 - central_res
run_str = '-'.join([str(site - run_start), str(site + run_end)])
chain_run_selection = ResidueIndexSelector(run_str)
return chain_run_selection
# Shouldn't happen, but just in case...
else:
print("Bug detected: received one input, not zero or two.")
exit()
def select_residues(design_peptide=False):
""" Makes residue selectors for HCV protease sections """
residue_selectors = {}
catalytic = ResidueIndexSelector('72,96,154')
residue_selectors['catalytic'] = catalytic
protease = ChainSelector("A")
residue_selectors['protease'] = protease
peptide = ChainSelector("B")
residue_selectors['peptide'] = peptide
mutable = mutable_residues_selector(design_peptide)
residue_selectors['mutable'] = mutable
packable = packable_residues_selector(mutable)
residue_selectors['packable'] = packable
other = other_residues_selector(mutable, packable)
residue_selectors['other'] = other
mm_pack = movemap_packable_residues_selector(peptide, mutable, packable)
residue_selectors['movemap_pack'] = mm_pack
return residue_selectors
def task_factory_builder(repacking_residues='all', designing_residues=None, \
extra_rotamers=[1,2]):
#Setting this up for now, need to implement these conditions:
#Repackable set, Designable Set, prevent ligand repacking
#Building a Task Factory for repacking/design around the neighbors residues
#Task factory builder should accept: repackable residues, designable
#residues, specifically preventing repacking residues, and prevent the rest.
#tf = task_factory_builder(repack=repacking_neighbors)
#distance - distance for Neighborhood residue selector - need to remov
print_out("Building task factory")
#Start up Task factory, starting ex1 and ex2, and include current
tf = TaskFactory()
tf.push_back(IncludeCurrent())
if extra_rotamers:
for x in extra_rotamers:
tf.push_back(ExtraRotamers(0, x, 1))
print(repacking_residues)
#Set up repack and prevent repack
repack = RestrictToRepackingRLT()
prevent = PreventRepackingRLT()
prevent_everything_else = NotResidueSelector()
residues_for_repack = None
residues_for_design = None
changing_residues = OrResidueSelector()
if repacking_residues != 'all':
print_out("Repacking these residues: " + str(repacking_residues))
residues_for_repack = ResidueIndexSelector(
','.join(repacking_residues))
changing_residues.add_residue_selector(residues_for_repack)
tf.push_back(OperateOnResidueSubset(repack, residues_for_repack))
if designing_residues:
residues_for_design = ResidueIndexSelector(designing_residues)
changing_residues.add_residue_selector(residues_for_design)
prevent_everything_else.set_residue_selector(changing_residues)
if repack == 'all':
tf.push_back(OperateOnResidueSubset(repack, prevent_everything_else))
else:
tf.push_back(OperateOnResidueSubset(prevent, prevent_everything_else))
return tf
def load_selections_from_json_file(filename, pose):
"""
Reads a selections.JSON file and transforms the selection into actual
ResidueSelectors, if the residues exist. Otherwise, save that selection as
None.
"""
with open(filename, "r") as file_in:
selections = json.load(file_in)
for selection in selections:
sel = ResidueIndexSelector(selections[selection])
try:
sel.apply(pose)
selections[selection] = sel
except RuntimeError:
selections[selection] = None
return selections
def apply_hbond_constraints(pose, first_strand, second_strand):
"""
Adds H-bond constraints to a pose, given lists of residues in two beta
strands. Lists should be given as strings of residues, such as '215,217'
and '185,187' for Htra1 with the extended peptide, or '199-203' and
'170-174' for HCV.
"""
# Create selectors for residues in each set
hb_res_1 = ResidueIndexSelector(first_strand)
hb_res_2 = ResidueIndexSelector(second_strand)
# Set up constraints generation
hbcg = HydrogenBondConstraintGenerator()
hbcg.set_residue_selector1(hb_res_1)
hbcg.set_residue_selector2(hb_res_2)
# Apply, return pose
hbcg.apply(pose)
return pose
def create_coord_cst(ref_pose=None, free_res_indexes=None):
coord_cst_gen = CoordinateConstraintGenerator()
if ref_pose:
coord_cst_gen.set_reference_pose(ref_pose)
if free_res_indexes:
free_res_selector = ResidueIndexSelector(','.join(
str(free_res_index) for free_res_index in free_res_indexes))
coord_cst_gen.set_residue_selector(
NotResidueSelector(free_res_selector))
return coord_cst_gen
def make_point_mutant_task_factory(site_1,
change_1,
site_2=None,
change_2=None,
fast=True):
"""
Given a site integer and a 1-letter residue name (or two sites and two
residue names), generates a TaskFactory to repack all residues
in a pose with the altered sites repacking to the new sequence.
"""
# Initialize selection of all altered residues
modified_residues = OrResidueSelector() # Keep list of mobile residues
# Make TaskFactory to input changes
task_factory = TaskFactory()
task_factory.push_back(IncludeCurrent())
if not fast:
task_factory.push_back(ExtraRotamers(0, 1, 1))
task_factory.push_back(ExtraRotamers(0, 2, 1))
# Force packing to new residue for the mutated site
res_selection_1 = ResidueIndexSelector(str(site_1))
aa_force_1 = RestrictAbsentCanonicalAASRLT()
aa_force_1.aas_to_keep(change_1)
task_factory.push_back(OperateOnResidueSubset(aa_force_1, res_selection_1))
modified_residues.add_residue_selector(res_selection_1)
# Add second site if necessary
if site_2 and change_2:
res_selection_2 = ResidueIndexSelector(str(site_2))
aa_force_2 = RestrictAbsentCanonicalAASRLT()
aa_force_2.aas_to_keep(change_2)
task_factory.push_back(
OperateOnResidueSubset(aa_force_2, res_selection_2))
modified_residues.add_residue_selector(res_selection_2)
# Repack all other residues to accommodate substitutions
unchanged_residues = NotResidueSelector(modified_residues)
repack = RestrictToRepackingRLT()
task_factory.push_back(OperateOnResidueSubset(repack, unchanged_residues))
return task_factory
def mutable_residues_selector(design_peptide=False, single_pep_res=None):
"""
Selects the residues in a shell around the peptide using the
InterGroupInterfaceByVectorSelector residue selector
Presently hard coded for HCV protease.
Decided to design just around peptide, not pep + cat, and only the
six mutable residues in the peptide.
"""
# Selecting regions
protease = ChainSelector("A")
catalytic = ResidueIndexSelector('72,96,154')
if single_pep_res:
variable_pep_res = ResidueIndexSelector(str(single_pep_res))
else:
variable_pep_res = ResidueIndexSelector("198-203")
# Making positive residue selector
rs = InterGroupInterfaceByVectorSelector()
rs.group1_selector(protease) # Protease
rs.group2_selector(variable_pep_res) # Peptide recognition region
rs.nearby_atom_cut(6)
rs.vector_dist_cut(8)
# Excluding the catalytic residues
limit_selection = AndResidueSelector()
limit_selection.add_residue_selector(NotResidueSelector(catalytic))
limit_selection.add_residue_selector(rs)
# If the peptide sequence is mutable
if design_peptide:
return limit_selection
# If only the protease is designable
else:
# Setting up exclusion of catalytic and peptide residues
exclusive_selection = AndResidueSelector()
exclusive_selection.add_residue_selector(limit_selection)
exclusive_selection.add_residue_selector(ChainSelector("A"))
return exclusive_selection
def index_interface(pose, substrate_indices, d=10):
"""This function takes a pose and a number of interface/substrate to consider and returns interface indices. The
value k and pose are not used..."""
# Selection for neighbor residues
focus_res = ','.join([str(j) for j in substrate_indices])
focus_selector = ResidueIndexSelector(focus_res)
interface = NeighborhoodResidueSelector()
interface.set_focus_selector(focus_selector)
interface.set_distance(d)
interface.set_include_focus_in_subset(False)
interface_indices = selector_to_list(pose, interface)
interface_indices.sort()
return interface_indices
def packable_residues_selector(mutable_selector):
"""
Selects the shell of neighbor residues to repack
Presently hard coded for HCV protease.
"""
# Selecting regions
mutable = mutable_selector
not_mutable = NotResidueSelector(mutable_selector)
catalytic = ResidueIndexSelector('72,96,154')
peptide = ChainSelector('B')
pep_not_mutable = AndResidueSelector()
pep_not_mutable.add_residue_selector(peptide)
pep_not_mutable.add_residue_selector(not_mutable)
# Selecting residues near mutable shell
near_mutable = InterGroupInterfaceByVectorSelector()
near_mutable.group1_selector(not_mutable)
near_mutable.group2_selector(mutable)
near_mutable.nearby_atom_cut(4)
near_mutable.vector_dist_cut(4)
# Selecting residues near the peptide
near_pep = InterGroupInterfaceByVectorSelector()
near_pep.group1_selector(not_mutable) # Protease
near_pep.group2_selector(peptide) # Peptide recognition region
near_pep.nearby_atom_cut(10)
near_pep.vector_dist_cut(12)
# Combining selections for near peptide and near mutable residues
wide_set = OrResidueSelector()
wide_set.add_residue_selector(near_mutable)
wide_set.add_residue_selector(near_pep)
# Setting up exclusion of catalytic and mutable residues
limit_selection = AndResidueSelector()
limit_selection.add_residue_selector(NotResidueSelector(catalytic))
limit_selection.add_residue_selector(wide_set)
limit_selection.add_residue_selector(not_mutable)
# Add back in the peptide
expand_selection = OrResidueSelector()
expand_selection.add_residue_selector(limit_selection)
expand_selection.add_residue_selector(pep_not_mutable)
return expand_selection
def uncap(pose):
"""
Remove all termini conformations of a pose, switching the atoms to normal
in-chain conformations.
"""
from pyrosetta.rosetta.core.select.residue_selector import \
ResidueIndexSelector
# Identify all termini in the given pose
cap_indexes = ""
for residue in pose.residues:
if residue.is_terminus():
cap_indexes += "%d," % (residue.seqpos())
caps = ResidueIndexSelector(cap_indexes[:-1])
# Remove termini conformations for the selection in the given pose
uncap_selection(pose, caps)
def design(self, antibody=True, antigen=False, pack_only=False):
self.pose = self.native_pose.clone()
info = self.pose.pdb_info()
tf = TaskFactory()
tf.push_back(operation.InitializeFromCommandline())
epi_res = ''
for res in self.epitopes_pose:
epi_res += '{0},'.format(res)
epi_selector = ResidueIndexSelector(epi_res)
antibody_selector = ChainSelector()
vec = vector1_std_string()
for chain in self.antibody:
vec.append(chain)
vec.append(chain)
antibody_selector.set_chain_strings(vec)
interface_res_selector = InterGroupInterfaceByVectorSelector(
epi_selector, antibody_selector)
interface_antibody_selector = AndResidueSelector(
interface_res_selector, antibody_selector)
if pack_only: tf.push_back(operation.RestrictToRepacking())
else:
if antigen: design_selector = epi_selector
elif antibody: design_selector = interface_antibody_selector
# FIRST select designable residues and prevent everything else from design and packing
no_design_selector = NotResidueSelector(design_selector)
prevent_repacking_rlt = operation.PreventRepackingRLT()
#restrict_topack = operation.RestrictToRepackingRLT()
prevent_design = operation.OperateOnResidueSubset(
prevent_repacking_rlt, no_design_selector, False)
tf.push_back(prevent_design)
print(tf.create_task_and_apply_taskoperations(self.pose))
packer = pack_min.PackRotamersMover('ref2015')
packer.task_factory(tf)
packer.apply(self.pose)
def calculate_sequence(self):
""" Determine full sequence, and compare to WT if available """
# Get full sequence
sm = SequenceMetric()
self.full_sequence = sm.calculate(self.design)
# If WT is available, compare
#if self.wt:
# Check that lengths are the same
assert self.design.total_residue() == self.wt.total_residue()
for i in range(1, self.design.total_residue() + 1):
ris = ResidueIndexSelector(str(i))
sm.set_residue_selector(ris)
final_aa = sm.calculate(self.design)
orig_aa = sm.calculate(self.wt)
pdb_number = int(self.wt.pdb_info().pose2pdb(i).split()[0])
if final_aa != orig_aa:
self.mutated_residues[pdb_number] = [orig_aa, final_aa]
def get_pose_discrimination_scores(pdb):
fix_file(pdb)
nam = basename(pdb)
pose = pose_from_pdb(pdb)
# Enzdes constraints
cstm = AddOrRemoveMatchCsts()
cstm.set_cst_action(ADD_NEW)
cstm.apply(pose)
sf = create_score_function('ref2015_cst')
sf(pose)
#pep_res = selector_to_list(pose, ChainSelector('B'))
pep_res = selector_to_list(pose, ResidueIndexSelector('212-218'))
neigh = NeighborhoodResidueSelector()
neigh.set_distance(8)
neigh.set_focus_selector(ChainSelector('B'))
neigh.set_include_focus_in_subset(False)
prot_res = selector_to_list(pose, neigh)
pep_score = 0
for i in pep_res:
pep_score += pose.energies().residue_total_energies(i)[st.total_score]
prot_score = 0
for i in prot_res:
prot_score += pose.energies().residue_total_energies(i)[st.total_score]
cst_score = 0
for i in [st.atom_pair_constraint, st.coordinate_constraint,
st.angle_constraint, st.dihedral_constraint]:
cst_score += pose.energies().total_energies()[i]
discriminator = 1 * prot_score + 1 * pep_score + 3.5 * cst_score
print(nam, prot_score, pep_score, cst_score, discriminator)
return [nam, prot_score, pep_score, cst_score, discriminator]
def apply_constraints(pose):
"""
Applies enzdes constraints from the input CST file to a pose
Also applies coordinate constraints to the original substrate peptide
which is assumed correct from the PDB
"""
# Enzdes constraints
cstm = AddOrRemoveMatchCsts()
cstm.set_cst_action(ADD_NEW)
cstm.apply(pose)
# Coordinate constraints
cg = CoordinateConstraintGenerator()
cs = ResidueIndexSelector('213-216')
cg.set_residue_selector(cs)
ac = AddConstraints()
ac.add_generator(cg)
ac.apply(pose)
return
def apply_constraints(pose):
"""
Applies enzdes constraints form the input CST file to a pose
Also applies coordinate constraints to the substrate peptide, assumed to
be chain B
"""
# Enzdes constraints
cstm = AddOrRemoveMatchCsts()
cstm.set_cst_action(ADD_NEW)
cstm.apply(pose)
# Coordinate constraints
cg = CoordinateConstraintGenerator()
ors = OrResidueSelector()
ors.add_residue_selector(ChainSelector('A')) # Constrain main backbone
ors.add_residue_selector(
ResidueIndexSelector('113-115')) # Preserving original peptide
cg.set_residue_selector(ors)
ac = AddConstraints()
ac.add_generator(cg)
ac.apply(pose)
return
def residue_selection(selection_one, selection_two):
#For now, just an or selector to group up the residues.
first = ResidueIndexSelector(selection_one)
second = ResidueIndexSelector(selection_one)
both = OrResidueSelector(first, second)
return both
def calc_residue_interaction(pose, sf, res, int_target):
ris = ResidueIndexSelector(str(res))
inte = InteractionEnergyMetric()
inte.set_scorefunction(sf)
inte.set_residue_selectors(ris, int_target)
return inte.calculate(pose)
def make_residue_changes(pose, sf, subst_seq, subst_start, cat_res,
manual_muts):
"""
Applies substrate sequence changes and manual mutations to a given pose.
This is done through repacking, so unlike SimpleThreadingMover, the side
chains don't begin clashing. This means that the residue selectors will be
more accurate, and that design can begin without an initial relax step.
pose is a Rosetta pose
sf is a Rosetta scorefunction
subst_seq is a string (doesn't need to be uppercase)
subst_start is an integer corresponding to the first of a contiguous block
of residues to re-sequence
manual_muts is a list of two-member lists, of the following form:
[site, single-letter residue name]
"""
# Create dict of {res: AA} for changes to make
res_changes = {}
# Add manual mutations list
if manual_muts:
print("\nApplying point substitutions:")
for m in manual_muts:
res_changes[int(m[0])] = m[1].upper()
print(m[0], m[1].upper())
# Add substrate threading to list of res changes
print("\nInserting substrate sequence:\n{}".format(subst_seq))
subst_range = range(subst_start, subst_start + len(subst_seq))
for n, i in enumerate(subst_range):
res_changes[i] = subst_seq[n].upper()
# Make TaskFactory to input changes
mobile_residues = OrResidueSelector() # Keep list of mobile residues
tf = TaskFactory()
# Force packing to target residue for each desired change
for r, aa in res_changes.items():
res_selection = ResidueIndexSelector(str(r))
restriction = RestrictAbsentCanonicalAASRLT()
restriction.aas_to_keep(aa.upper())
tf.push_back(OperateOnResidueSubset(restriction, res_selection))
mobile_residues.add_residue_selector(res_selection)
# Repack nearby residues to accommodate substitutions
shell = NeighborhoodResidueSelector()
shell.set_focus_selector(mobile_residues)
shell.set_include_focus_in_subset(False)
shell.set_distance(8)
# Exclude catalytic residues
if cat_res:
catalytic = ResidueIndexSelector(','.join([str(i) for i in cat_res]))
not_catalytic = NotResidueSelector(catalytic)
shell = selector_intersection(shell, not_catalytic)
restrict = RestrictToRepackingRLT()
tf.push_back(OperateOnResidueSubset(restrict, shell))
# Prevent repacking of all other residues
unchanging = NotResidueSelector(OrResidueSelector(mobile_residues, shell))
prevent = PreventRepackingRLT()
tf.push_back(OperateOnResidueSubset(prevent, unchanging))
# Apply changes with PackRotamersMover
pt = tf.create_task_and_apply_taskoperations(pose)
prm = PackRotamersMover(sf, pt)
mutated_pose = Pose(pose)
prm.apply(mutated_pose)
return mutated_pose
def create_relax_task_factory(point_mutations: list = list(),
active_site_positions: list = list(),
neighborhood: float = 0):
task_factory = TaskFactory()
task_factory.push_back(IncludeCurrent())
repack = RestrictToRepackingRLT()
prevent = PreventRepackingRLT()
if len(point_mutations) > 0:
# Mutate
mutated_selector = OrResidueSelector()
for point_mutation in point_mutations:
mutation_info = point_mutation.split(',')
restriction = RestrictAbsentCanonicalAASRLT()
restriction.aas_to_keep(mutation_info[1])
point_mutation_selector = ResidueIndexSelector(mutation_info[0])
task_factory.push_back(
OperateOnResidueSubset(restriction, point_mutation_selector))
mutated_selector.add_residue_selector(point_mutation_selector)
# Repack and static
if neighborhood > 0:
if len(active_site_positions) > 0:
# Repack
enzyme_core_selector = ResidueIndexSelector(','.join(
str(active_site_position)
for active_site_position in active_site_positions))
designable_repacking_selector = NeighborhoodResidueSelector()
designable_repacking_selector.set_focus_selector(
OrResidueSelector(mutated_selector, enzyme_core_selector))
designable_repacking_selector.set_distance(neighborhood)
designable_repacking_selector.set_include_focus_in_subset(True)
repacking_selector = AndResidueSelector(
designable_repacking_selector,
NotResidueSelector(mutated_selector))
task_factory.push_back(
OperateOnResidueSubset(repack, repacking_selector))
# Static
task_factory.push_back(
OperateOnResidueSubset(prevent,
designable_repacking_selector,
True))
else:
# Repack
repacking_selector = NeighborhoodResidueSelector()
repacking_selector.set_focus_selector(mutated_selector)
repacking_selector.set_distance(args.neighborhood)
repacking_selector.set_include_focus_in_subset(False)
task_factory.push_back(
OperateOnResidueSubset(repack, repacking_selector))
# Static
mutated_and_repacking_selector = OrResidueSelector(
mutated_selector, repacking_selector)
task_factory.push_back(
OperateOnResidueSubset(prevent,
mutated_and_repacking_selector,
True))
else:
# Repack
task_factory.push_back(
OperateOnResidueSubset(repack, mutated_selector, True))
else:
if neighborhood > 0:
if len(active_site_positions) > 0:
# Repack
enzyme_core_selector = ResidueIndexSelector(','.join(
str(active_site_position)
for active_site_position in active_site_positions))
repacking_selector = NeighborhoodResidueSelector()
repacking_selector.set_focus_selector(enzyme_core_selector)
repacking_selector.set_distance(neighborhood)
repacking_selector.set_include_focus_in_subset(True)
task_factory.push_back(
OperateOnResidueSubset(repack, repacking_selector))
# Static
task_factory.push_back(
OperateOnResidueSubset(prevent, repacking_selector, True))
else:
# Static
task_factory.push_back(PreventRepacking())
else:
# Repack
task_factory.push_back(RestrictToRepacking())
return task_factory
def get_cat_res(in_pose):
"""
For an input pose (presently assumed to be a serine protease), several
steps are performed to determine the catalytic triad residues.
Firstly, the pose is stripped down to the largest chain, which should be
the main. This is to eliminate substrate chains that might falsly appear to
be part of the triad.
Secondly, selections of ser, his, and acid (asp and glu) residues are
collected.
Thirdly, a crude screen is conducted to weed out obviously false residues.
The set of histidines is collected and scanned through individually. For
each histidine, the set of neighboring residues within 10 A is checked for
overlap with the sets of serines and acids, and if either set intersection
is empty, which means that there is no nearby serine and/or acid, then that
histidine cannot be part of the triad and is eliminated. Remaining
histidines are collected with the lists of their potential ser and acid
counterparts.
Fourthly, the geometry of potential triads are examined. The atom distances
are checked between the ND's of the histidine and the respective CB of
serine and OD of the acid residue. The ND-CB distance should be 3.3 +/-
0.5 A, and the ND-OD distance should be 2.6 +/- 0.5 A.
"""
# 1. Isolate the protease main chain (no substrate)
pose = isolate_main_chain(in_pose)
# 2. Get selectors of serines, histidines, and acidic residues (aspartic and glutamic)
target_res = ['SER', 'HIS', 'ASP,GLU']
name_selectors = {}
for tr in target_res:
# Need to set name3 in selectors to get terminal residues, HIS_D, etc.
rns = ResidueNameSelector()
rns.set_residue_name3(tr)
name_selectors[tr] = rns
# 3. Scan through list of histidines to weed out any that don't have nearby
# serine and acid residues within 10 A
potential_his = selector_to_list(pose, name_selectors['HIS'])
potential_triads = {}
for h in potential_his:
# Make selector for all residues with CA within 10 A, excluding self
nrs = NeighborhoodResidueSelector(ResidueIndexSelector(str(h)), 10)
nrs.set_include_focus_in_subset(False)
# Select intersection of neighbors and serines
no_match, potential_ser = \
check_selector_nonzero_intersections([nrs, name_selectors['SER']], pose)
if no_match:
# If there are no nearby serines, this can't be catalytic hist
continue
# Select intersection of neighbors and acid residues
no_match, potential_ac = \
check_selector_nonzero_intersections([nrs, name_selectors['ASP,GLU']], pose)
if no_match:
# If there are no nearby acid residues, this can't be catalytic hist
continue
# Histidine has both requisite neighbors. Collect that residue and the
# lists of neighbors.
potential_triads[h] = [potential_ser, potential_ac]
# 4. Examine rotamer geometry of potential triads
for h, [sers, acids] in potential_triads.items():
for s in sers:
for a in acids:
pass
add_csts.apply(pose)
# constraint
if args.constraints:
add_fa_constraints_from_cmdline(pose, sfxn)
# enzyme design constraint
if args.enzyme_design_constraints:
enzdes_cst = create_enzdes_cst()
enzdes_cst.apply(pose)
# declare movers
movers = list()
if not args.rmsd:
rmsdm = RMSDMetric(pose)
else:
rmsdm = RMSDMetric(
pose,
ResidueIndexSelector(','.join(str(res) for res in args.rmsd)))
movers.append(rmsdm)
if args.fast_relax:
tf = create_relax_task_factory(point_mutations = args.mutations, active_site_positions = active_site_positions, \
neighborhood = args.neighborhood)
elif args.designable_sites or args.substrates or args.enzyme_design_constraints:
if args.favor_native_residue:
favor_nataa = FavorNativeResidue(pose, args.favor_native_residue)
if args.designable_sites:
tf = create_design_task_factory(args.designable_sites, active_site_positions = active_site_positions, \
neighborhood = args.neighborhood, cystine = args.cystine, noncanonical_amino_acids = args.noncanonical_amino_acids)
else:
tf = create_enzyme_design_task_factory(active_site_positions, neighborhood = args.neighborhood, \
cystine = args.cystine, noncanonical_amino_acids = args.noncanonical_amino_acids)
else:
tf = create_relax_task_factory(point_mutations=args.mutations,
prem.set_scorefunction(sf)
tem = TotalEnergyMetric()
tem.set_scorefunction(sf)
results = []
for p in pdbs:
pp = pose_from_pdb(p)
results.append([p, tem.calculate(pp), prem.calculate(pp)])
s = prem.calculate(pose)
with open('res_results.txt', 'w') as w:
for r in results:
res_scores = []
for i in range(1, 116):
res_scores.append(r[2][i])
ol = [r[0], r[1]] + res_scores
w.write(','.join([str(x) for x in ol]) + '\n')
csb = ChainSelector('B')
with open('res_interactions.csv', 'w') as w:
for p in pdbs:
pp = pose_from_pdb(p)
tot_e = tem.calculate(pp)
this_res = [p, tot_e]
for i in range(1, 115):
ris = ResidueIndexSelector(str(i))
inte = InteractionEnergyMetric()
inte.set_scorefunction(sf)
inte.set_residue_selectors(ris, csb)
this_res.append(inte.calculate(pp))
w.write(','.join([str(x) for x in this_res]) + '\n')
def main(args):
#Determining if the ligand should be removed or not
#Need to be moved here as checking if the lig remains is
#Imperative for how the CST file is dealt with.
out_ligand_file = 'yeslig'
if args.remove_ligand:
args.rigid_ligand = False
out_ligand_file = 'nolig'
params = [args.unnatural]
uaa = params[0].split('/')[-1].strip(".params")
if args.ligand_type == 'ligand':
params.append(args.ligand_name)
lig_name = args.ligand_name.split('/')[-1].strip('.params')
init_args = ' '.join(['-run:preserve_header', '-extra_res_fa'] + params)
#Adding enzdes constraint file - and editing it - if necessary
if args.enzdes_constraint_file:
if out_ligand_file == 'nolig':
if args.ligand_type == 'protein':
lig_search = args.residue_set
elif args.ligand_type == 'ligand':
lig_search = args.ligand_name
args.input_pdb, args.enzdes_constraint_file = enzdes_constraint_eliminator(\
args.input_pdb, args.enzdes_constraint_file, ligand=lig_search )
#init_args = init_args + " -enzdes::cstfile " + str(args.enzdes_constraint_file)
#Starting up rosetta with the appropriate params files -
#-- need both unnatural aa and ligand file to be added (if the ligand isn't a protein).
print_out("emd182::Starting rosetta with the following parameters: " +
init_args)
pr.init(init_args)
file_options = pr.rosetta.core.io.StructFileRepOptions()
file_options.set_preserve_header(bool(True))
pose = pr.pose_from_pdb(args.input_pdb)
#Residue selection
if args.residue_number != '0':
delta_resi = ResidueIndexSelector(args.residue_number)
#Determining if the interacting ligand is a protein or small molecule
#and selecting the appropriate residues
if args.ligand_type == 'protein':
ligand = ResidueIndexSelector(args.residue_set)
elif args.ligand_type == 'ligand':
ligand = ResidueNameSelector()
ligand.set_residue_name3(lig_name)
#Adding the unnatural at the mutation site
print_out("Loading Unnatural " + uaa + \
" onto residue number " + args.residue_number )
if args.residue_number in selector_to_vector(ligand, pose):
print_out(
"Selected residue number IS a part of the ligand. Don't do that. \
Chosen residue number: " + str(args.residue_number))
#Setting up Mutation function on the correct residue, so long as the
#residue isn't #0
if args.residue_number != '0':
mutater = pr.rosetta.protocols.simple_moves.MutateResidue()
mutating_residue = ResidueIndexSelector(args.residue_number)
mutater.set_selector(mutating_residue)
mutater.set_res_name(uaa)
print_out("emd182::loading residue to mutate into: " +
str(args.residue_number))
lig_vector = selector_to_vector(ligand, pose)
if args.residue_number != '0':
if args.ligand_type == 'protein' and delta_resi in lig_vector:
print_out("Selected residue for mutation is part of input selection: " + \
delta_resi + " is selected, but is contained in " \
+ str(lig_vector))
print_out("Exiting the python script")
sys.exit()
print_out("emd182::Start mutations and relaxation script " \
+ str(args.nstruct) + " times.")
if args.residue_number == '0':
args.nstruct = 1
for struct in range(0, args.nstruct):
print_out(struct)
mutant_pose = pr.Pose(pose)
#Residue selection
if args.residue_number != '0':
delta_resi = ResidueIndexSelector(args.residue_number)
#apply mutation
if args.residue_number != '0':
mutater.apply(mutant_pose)
if args.dihedral_constraint_atoms:
dihedral_atoms = args.dihedral_constraint_atoms.split(';')
dihedral_values = args.dihedral_constraint_degrees.split(';')
dihedral_cst_stdev = args.dihedral_constraint_stdev.split(';')
for dihedrals in range(len(dihedral_atoms)):
apply_dihedral_constraint(dihedral_atoms[dihedrals].split(','), \
delta_resi, mutant_pose, dihedral_values[dihedrals],\
dihedral_cst_stdev[dihedrals])
move_map = build_move_map(True, True, True)
#Scoring the pose
sf = pr.rosetta.core.scoring.ScoreFunction()
if args.symmdef_file:
#Setting a residueindexselector to eliminate all extra aa post design
pre_symm_ris = ResidueIndexSelector()
all_aa = []
for i in range(1, mutant_pose.total_residue() + 1):
all_aa.append(''.join(
mutant_pose.pdb_info().pose2pdb(i).split()))
pre_symm_ris.set_index(','.join(all_aa))
#sys.exit()
#Symmetrizing
sfsm = SetupForSymmetryMover(args.symmdef_file)
sfsm.apply(mutant_pose)
sf = pr.rosetta.core.scoring.symmetry.SymmetricScoreFunction()
sf.add_weights_from_file('ref2015_cst')
else:
sf = pr.rosetta.core.scoring.ScoreFunction()
sf.add_weights_from_file('ref2015_cst')
sf(mutant_pose)
#apply mutation
if args.residue_number != '0':
mutater.apply(mutant_pose)
if args.dihedral_constraint_atoms:
dihedral_atoms = args.dihedral_constraint_atoms.split(';')
dihedral_values = args.dihedral_constraint_degrees.split(';')
dihedral_cst_stdev = args.dihedral_constraint_stdev.split(';')
for dihedrals in range(len(dihedral_atoms)):
apply_dihedral_constraint(dihedral_atoms[dihedrals].split(','), \
delta_resi, mutant_pose, dihedral_values[dihedrals],\
dihedral_cst_stdev[dihedrals])
#Making appropriate residue selections base on if the ligand will be
#Removed or not, as well as if the ligand is rigid or not.
residues_around_ligand = ligand_neighbor_selection(ligand, args.radius, \
mutant_pose, bool(args.rigid_ligand) )
if args.residue_number != '0':
residues_around_mutant = ligand_neighbor_selection(mutating_residue, \
args.radius, mutant_pose, True)
design_around_mutant = None
#if design is turned on, will design around the mutant.
#Can add more variability later
if args.design and args.residue_number != '0':
designing_residues = ligand_neighbor_selection(mutating_residue, \
args.design, mutant_pose, False)
#Combining the repacking neighborhoods around the ligand and mutant
if args.residue_number != '0':
repacking_neighbors = residue_selection(residues_around_ligand, \
residues_around_mutant)
else:
repacking_neighbors = ResidueIndexSelector(residues_around_ligand)
#Specifically converting residue selectors to vectors, and removing the
#Appropriate residue ids from the lists of repacking or designing residues
repacking_resids = selector_to_vector(repacking_neighbors, mutant_pose)
if args.rigid_ligand:
lig_resids = selector_to_vector(ligand, mutant_pose)
for res in lig_resids:
if res in repacking_resids:
repacking_resids.remove(res)
if args.design:
designing_resids = selector_to_vector(designing_residues,
mutant_pose)
for res in lig_resids:
if res in designing_resids:
repacking_resids.remove(res)
#If the remove-ligand is called, will remove the ligand from the mutant pose
#Change 'remove-ligand' to 'translate region - makes for
if args.remove_ligand:
used_xyzs = []
for chain in args.remove_ligand.split(','):
x, y, z = random_direction_selector(used_xyzs)
used_xyzs.append(np.array([x, y, z]))
trans_vec = pr.rosetta.numeric.xyzVector_double_t(x, y, z)
trans = pr.rosetta.protocols.rigid.RigidBodyTransMover(
trans_vec)
jump_id = pr.rosetta.core.pose.get_jump_id_from_chain(
chain, mutant_pose)
trans.rb_jump(jump_id)
trans.step_size(500.0)
trans.apply(mutant_pose)
tf = task_factory_builder(repacking_residues=repacking_resids, \
designing_residues=design_around_mutant)
move_map = build_move_map(True, True, True)
#turn on match constraints if needed:
if args.enzdes_constraint_file:
if not args.remove_ligand:
print_out(
'The code will break if the constraint file is attached \
to the ligand that is being removed')
apply_match_constraints(mutant_pose, args.enzdes_constraint_file)
print('just checking')
#turn off constraints if requested - default is on
if not args.no_constraints:
mutant_pose = coord_constrain_pose(mutant_pose)
#Repack or minimize if selected
if args.fast_relax:
print_out("Running With Fast Relax")
new_pose = fast_relax_mutant(mutant_pose, tf, move_map, sf)
else:
print_out("Running repack and minimize")
new_pose = repack_and_minimize_mutant(mutant_pose, tf, move_map,
sf)
#output the name of the file
#Includes original pdb namne, residue number and uaa, nstruct number,
#and if the ligand is included or not.
base_pdb_filename = args.input_pdb.split('/')[-1].split(
'_')[0].replace('.pdb', '')
outname = '{}_{}_{}_{}_{}.pdb'.format(base_pdb_filename, \
args.residue_number, uaa, out_ligand_file, str(struct))
if args.residue_number == '0':
outname = '{}_{}_min.pdb'.format(base_pdb_filename, \
out_ligand_file)
if args.out_suffix:
outname = outname.replace(".pdb", args.out_suffix + ".pdb")
print_out("Outputting protein file as : " + outname)
out_dir = out_directory(args.out_directory)
outname = '/'.join([out_dir, outname])
print_out("Writing Outputting protein file as : " + outname)
#Symmetry needs to be broken to load the proteins in another
#script to analyze them symmetrically.
if args.symmdef_file:
full_out = outname.replace('.pdb', 'sym.pdb')
new_pose.dump_pdb(full_out)
not_symm = NotResidueSelector(pre_symm_ris)
remove_symmetrized_aa = DeleteRegionMover()
remove_symmetrized_aa.set_residue_selector(not_symm)
remove_symmetrized_aa.apply(new_pose)
new_pose.dump_pdb(outname)
else:
new_pose.dump_pdb(outname)