def Model_Mutations_old(pdbfile,
mol2files,
mutations,
max_overlap=0.5,
return_score=False):
"""Model a number of mutations in a pdbfile when one or more ligands are present"""
#
# Initialise mutate routines
#
MUT = Mutate(max_bump=max_overlap)
#
# Read PDB file
#
import Protool
P = Protool.structureIO()
P.readpdb(pdbfile)
P.remove_all_hydrogens()
#
# Read mol2 file
#
L = Protool.ligand(P)
for mol2file in mol2files:
print "Added %s with tag 'LIGAND'" % mol2file
L.readmol2(mol2file, tag='LIGAND')
#
# Pass combined pdb file to mutate routines and mutate
#
MUT.new_PDB(P)
import pKa.pKD_tools as pKD_tools
total_bump = 0.0
#
# Model
#
for mutation in mutations:
#
# Get info
#
resid = pKD_tools.get_resid_from_mut(mutation)
newres = pKD_tools.get_newrestyp_from_mut(mutation)
oldres = pKD_tools.get_oldrestyp_from_mut(mutation)
bump_score = MUT.mutate(resid, newres, orgtype=oldres)
if bump_score is None or bump_score is False or bump_score > max_overlap:
print 'Cannot model this set of mutations - too many bumps'
return False, 20.0
print 'Bump score for %s: %5.3f' % (mutation, bump_score)
total_bump = total_bump + bump_score
print 'Total bump score for all mutations: %5.3f' % (bump_score)
if return_score:
return MUT, bump_score
return MUT
def Model_Mutations_old(pdbfile,mol2files,mutations,max_overlap=0.5,return_score=False):
"""Model a number of mutations in a pdbfile when one or more ligands are present"""
#
# Initialise mutate routines
#
MUT=Mutate(max_bump=max_overlap)
#
# Read PDB file
#
import Protool
P=Protool.structureIO()
P.readpdb(pdbfile)
P.remove_all_hydrogens()
#
# Read mol2 file
#
L=Protool.ligand(P)
for mol2file in mol2files:
print "Added %s with tag 'LIGAND'" %mol2file
L.readmol2(mol2file,tag='LIGAND')
#
# Pass combined pdb file to mutate routines and mutate
#
MUT.new_PDB(P)
import pKa.pKD_tools as pKD_tools
total_bump=0.0
#
# Model
#
for mutation in mutations:
#
# Get info
#
resid=pKD_tools.get_resid_from_mut(mutation)
newres=pKD_tools.get_newrestyp_from_mut(mutation)
oldres=pKD_tools.get_oldrestyp_from_mut(mutation)
bump_score=MUT.mutate(resid,newres,orgtype=oldres)
if bump_score is None or bump_score is False or bump_score>max_overlap:
print 'Cannot model this set of mutations - too many bumps'
return False,20.0
print 'Bump score for %s: %5.3f' %(mutation,bump_score)
total_bump=total_bump+bump_score
print 'Total bump score for all mutations: %5.3f' %(bump_score)
if return_score:
return MUT,bump_score
return MUT
def Model_Mutations(pdbfile,
mol2files,
mutations,
max_overlap=0.5,
max_totalbump=1.0,
return_score=False,
store_mutation_operations=False):
"""Model a number of mutations in a pdbfile when one or more ligands are present"""
#
# Check for stupidity
#
if max_overlap > max_totalbump:
max_totalbump = max_overlap
print 'Adjusted total bump cutoff to %5.2f' % max_totalbump
#
# Read PDB file
#
import Protool
P = Protool.structureIO()
P.readpdb(pdbfile)
P.remove_all_hydrogens()
#
# Read mol2 file
#
L = Protool.ligand(P)
for mol2file in mol2files:
print "Added %s with tag 'LIGAND'" % mol2file
L.readmol2(mol2file, tag='LIGAND')
#
# Get the pdb lines
#
pdblines = P.writepdb('junk.pdb', nowrite=True)
#
# Pass the lines to FFF
#
import FFF.FFFcontrol
myFFF = FFF.FFFcontrol.FFF()
myFFF.parse_lines(pdblines)
#myFFF.soup_stat()
Model = FFF.FFFcontrol.model_class(myFFF, Rotamerlib, FFFaadef_dir)
#
# Store the wild type PDB file
#
if store_mutation_operations:
wt_lines = myFFF.make_pdblines('PDB')
#
import pKa.pKD_tools as pKD_tools
total_bump = 0.0
for mutation in mutations:
resid = pKD_tools.get_resid_from_mut(mutation)
chainid = resid.split(':')[0]
resid = resid.split(':')[1]
#
# Get rid of the leading zeros
#
done = False
while not done:
if resid[0] == '0' and len(resid) > 1:
resid = resid[1:]
else:
done = True
#
newres = pKD_tools.get_newrestyp_from_mut(mutation)
oldres = pKD_tools.get_oldrestyp_from_mut(mutation)
opttype = 3 # Rotamer library
energies = Model.Mutate(chainid, resid, newres, opttype, max_overlap)
bump_score = energies[0]
Sum = energies[1]
Coulomb = energies[2]
#
total_bump = total_bump + bump_score
print 'Bump score: %5.2f, total bump: %5.2f' % (bump_score, total_bump)
if bump_score > max_overlap or total_bump > max_totalbump:
print 'Cannot model this set of mutations - too many bumps'
if return_score:
return False, total_bump
else:
return False
print 'Bump score for %s: %5.3f' % (mutation, bump_score)
print 'Total bump score for all mutations: %5.3f' % (total_bump)
class FFF_fix:
def __init__(self, FFF):
self.PI = FFF
return
#
# Create the instance
#
FFF_instance = FFF_fix(myFFF)
#
# Keep track of the changes that were made to the PDB file
#
if store_mutation_operations:
mut_lines = FFF_instance.PI.make_pdblines('PDB')
import Protool
WT = Protool.structureIO()
WT.parsepdb(wt_lines)
wt_atoms = sorted(WT.atoms.keys())
#
MUT = Protool.structureIO()
MUT.parsepdb(mut_lines)
#
mut_atoms = sorted(MUT.atoms.keys())
wt_count = 0
mutcount = 0
def coord_diff(atom1, atom2):
diff = 0.0
for coord in ['X', 'Y', 'Z']:
diff = diff + abs(atom1[coord] - atom2[coord])
return diff
operations = []
for atom in wt_atoms:
if not atom in mut_atoms:
operations.append(['delete', atom, WT.atoms[atom]])
elif coord_diff(WT.atoms[atom], MUT.atoms[atom]) > 0.1:
operations.append(['delete', atom, WT.atoms[atom]])
operations.append(['add', atom, MUT.atoms[atom]])
else:
pass
for atom in mut_atoms:
if not atom in wt_atoms:
operations.append(['add', atom, MUT.atoms[atom]])
#
# Store these in FFF_fix
#
FFF_instance.mutate_operations = operations[:]
#
# Return the info
#
if (return_score):
return FFF_instance, total_bump
return FFF_instance
def checkModels(self, DB=None, callback=None, selected=None, usemutationcodes=False):
"""Check that we have modelled a structure for everything we can"""
if DB == None:
return
proteins = DB.getRecs()
refprot = DB.meta.refprotein
refseq = DB[refprot].aaseq
refaa = self.AAList2String(refseq)
refpdb = DB[refprot].Structure
refpdbfile = os.path.join(os.getcwd(), 'ref.pdb')
self.writePDB(refpdb, refpdbfile)
failed = []
# Check that Protool is loaded
if not self.MUT:
self.initProtool()
#Create protool oinstance for ref pdb
import Protool
Xref = Protool.structureIO()
Xref.parsepdb(refpdb)
# Find all potential parents
records_with_structure=[]
for protein in proteins:
rec = DB.get(protein)
if rec.hasStructure() == 'available':
records_with_structure.append(protein)
# Loop over selected or all
if selected == None:
selected = list(set(proteins) - set(records_with_structure))
numrecords=len(selected)
count=1
for protein in selected:
rec = DB.get(protein)
if rec.hasStructure() == 'available':
continue
print 'Protein:', protein
#if no sequence try create one from mutation code
if rec.aaseq == None and rec.Mutations != None:
#print refaa
print 'no sequence, using mutation code and ref protein seq'
import PEATSA.Core as Core
print 'Record has mutation code %s' %rec.Mutations
mutationSet = Core.Data.MutationSet(rec.Mutations)
mutseq = mutationSet.applyToSequence(refaa, id='A', pdb=Xref)
rec.aaseq = self.string2AAseq(mutseq)
parent_with_structure = []
for parent in records_with_structure:
parentrec = DB.get(parent)
is_parent, operations = rec.getAncestry(parentrec)
# We can only model on X-ray structures
if parentrec.hasStructure() == 'available' and is_parent:
parent_with_structure.append([parent, len(operations)])
# Record failure to model
if parent_with_structure == []:
continue
# Find the best parent
def compare_func(x,y):
if x[1]>y[1]:
return 1
elif x[1]==y[1]:
return 0
if x[1]<y[1]:
return -1
parent_with_structure.sort(cmp=compare_func)
parent = parent_with_structure[0][0]
operations = rec.getAncestry(parentrec)[1]
print 'Using %s as template with %d operations.' %(parent, len(operations))
# Start the modelling
pdblines = parentrec.Structure
# Load the pdb file
import Protool
X=Protool.structureIO()
X.parsepdb(pdblines)
self.MUT.new_PDB(X)
self.MUT.max_tolerated_bump=0.5
atom_changes=[]
skip_protein=None
self.MUT.new_mutation()
for operation in operations:
# Is this a deletion?
if operation.find('delete')!=-1:
print 'This is a deletion - Jens should write code for modelling this'
print 'Deletion ignored for now'
continue
elif operation.find('insert')!=-1:
print 'This is an insertion - Jens should write code for modelling insertions'
print 'Insertion ignored for now'
continue
# This is a normal mutation
# Get the residue number, old residue and new residue
import pKa.pKD_tools as pKD_tools
new_res = pKD_tools.get_newrestyp_from_mut(operation)
old_res = pKD_tools.get_oldrestyp_from_mut(operation)
resid = pKD_tools.get_resid_from_mut(operation)
#print operation, resid
if not X.residues.has_key(resid):
print 'No structural info for mutation %8s. Not modelling this mutation\n' %operation
skip_protein=True
continue
# Actually make the mutation
bump_score=self.MUT.mutate(resid,new_res,orgtype=old_res)
print 'Mutation: %s, bump_score: %s' %(resid+new_res,str(bump_score))
if bump_score is None:
skip_protein=True
break
else:
atom_changes=atom_changes+self.MUT.mutate_operations
self.MUT.mutate_operations=[]
# Update progress
completion = float(count)/float(numrecords)*100.0
if callback != None:
callback(completion)
else:
print '%4d of %4d, completion; %5.2f%%' %(count,float(numrecords),completion)
count=count+1
# Did it work?
if skip_protein:
print
print 'Modelling failed for %s' %protein
failed.append(protein)
rec.Structure = 'Bumps'
rec.structuretype = 'failed model'
continue
# We have all sets of changes in atom_changes
rec.Structure = {'Rotamer_operations': atom_changes}
rec.Structure['parent'] = parent
rec.structuretype = 'peat model'
print 'Done'
if len(failed)>0:
print 'Failed to model the following proteins:'
for f in failed: print f
return
def makemutantSequence(self, sequence, operations):
"""Apply the specified mutations to a sequence and return the mutant seq
Sequence must be in the [[A:0001:ALA],[A:0002:GLU]] format
Operations is a list of the following types:
Mutations: A:0001:ALA:ASP
Deletions: delete:A:0002:GLU
Insertions: insert:1:A:0003:THR:ALA, insert:2:A:0003:THR:TRP (insert THR,TRP after A:0003:THR)
Operations are always performed in sequence numbering order """
if operations==[]:
return sequence
ops_sorted={}
insertions=[]
for operation in operations:
s_op=operation.split(':')
# Normal mutation
import pKa.pKD_tools as pKD_tools
resid=pKD_tools.get_resid_from_mut(operation)
if ops_sorted.has_key(resid):
raise Exception('More than one operation on the same residue: %s' %resid)
ops_sorted[resid]=['mutate',operation]
# Perform the operations
new_seq=[]
new_count=None
new_chain=None
for resid,restyp in sequence:
# Make sure that the chain hasn't changed or if we are at the beginning then init
if resid.split(':')[0]!=new_chain:
#Initialise
sp_resid=resid.split(':')
new_chain=sp_resid[0]
new_count=int(sp_resid[1])
newresid='%s:%s' %(new_chain,string.zfill(new_count,4))
# Does this residue have an operation?
if ops_sorted.has_key(resid):
op=ops_sorted[resid]
if op[0]=='delete':
# Deletion
if op[1]==restyp:
pass # This deletes the residue
else:
raise Exception('Incorrect org residue in deletion: %s' %op)
elif op[0]=='insert':
# Insertion
inserts=op[1].keys()
inserts.sort()
for i in inserts:
if i[0]==restyp:
new_seq.append([newresid,i[1]])
new_count=new_count+1
newresid='%s:%s' %(new_chain,string.zfill(new_count,4))
elif op[0]=='mutate':
# Mutation
import pKa.pKD_tools as pKD_tools
orgres=pKD_tools.get_oldrestyp_from_mut(op[1])
if orgres==restyp:
new_seq.append([newresid,pKD_tools.get_newrestyp_from_mut(op[1])])
new_count=new_count+1
newresid='%s:%s' %(new_chain,string.zfill(new_count,4))
pass
else:
raise Exception('Unknown mutations spec: %s' %op)
else:
new_seq.append([resid,restyp])
new_count=new_count+1
newresid='%s:%s' %(new_chain,string.zfill(new_count,4))
return new_seq
def checkModels(self,
DB=None,
callback=None,
selected=None,
usemutationcodes=False):
"""Check that we have modelled a structure for everything we can"""
if DB == None:
return
proteins = DB.getRecs()
refprot = DB.meta.refprotein
refseq = DB[refprot].aaseq
refaa = self.AAList2String(refseq)
refpdb = DB[refprot].Structure
refpdbfile = os.path.join(os.getcwd(), 'ref.pdb')
self.writePDB(refpdb, refpdbfile)
failed = []
# Check that Protool is loaded
if not self.MUT:
self.initProtool()
#Create protool oinstance for ref pdb
import Protool
Xref = Protool.structureIO()
Xref.parsepdb(refpdb)
# Find all potential parents
records_with_structure = []
for protein in proteins:
rec = DB.get(protein)
if rec.hasStructure() == 'available':
records_with_structure.append(protein)
# Loop over selected or all
if selected == None:
selected = list(set(proteins) - set(records_with_structure))
numrecords = len(selected)
count = 1
for protein in selected:
rec = DB.get(protein)
if rec.hasStructure() == 'available':
continue
print 'Protein:', protein
#if no sequence try create one from mutation code
if rec.aaseq == None and rec.Mutations != None:
#print refaa
print 'no sequence, using mutation code and ref protein seq'
import PEATSA.Core as Core
print 'Record has mutation code %s' % rec.Mutations
mutationSet = Core.Data.MutationSet(rec.Mutations)
mutseq = mutationSet.applyToSequence(refaa, id='A', pdb=Xref)
rec.aaseq = self.string2AAseq(mutseq)
parent_with_structure = []
for parent in records_with_structure:
parentrec = DB.get(parent)
is_parent, operations = rec.getAncestry(parentrec)
# We can only model on X-ray structures
if parentrec.hasStructure() == 'available' and is_parent:
parent_with_structure.append([parent, len(operations)])
# Record failure to model
if parent_with_structure == []:
continue
# Find the best parent
def compare_func(x, y):
if x[1] > y[1]:
return 1
elif x[1] == y[1]:
return 0
if x[1] < y[1]:
return -1
parent_with_structure.sort(cmp=compare_func)
parent = parent_with_structure[0][0]
operations = rec.getAncestry(parentrec)[1]
print 'Using %s as template with %d operations.' % (
parent, len(operations))
# Start the modelling
pdblines = parentrec.Structure
# Load the pdb file
import Protool
X = Protool.structureIO()
X.parsepdb(pdblines)
self.MUT.new_PDB(X)
self.MUT.max_tolerated_bump = 0.5
atom_changes = []
skip_protein = None
self.MUT.new_mutation()
for operation in operations:
# Is this a deletion?
if operation.find('delete') != -1:
print 'This is a deletion - Jens should write code for modelling this'
print 'Deletion ignored for now'
continue
elif operation.find('insert') != -1:
print 'This is an insertion - Jens should write code for modelling insertions'
print 'Insertion ignored for now'
continue
# This is a normal mutation
# Get the residue number, old residue and new residue
import pKa.pKD_tools as pKD_tools
new_res = pKD_tools.get_newrestyp_from_mut(operation)
old_res = pKD_tools.get_oldrestyp_from_mut(operation)
resid = pKD_tools.get_resid_from_mut(operation)
#print operation, resid
if not X.residues.has_key(resid):
print 'No structural info for mutation %8s. Not modelling this mutation\n' % operation
skip_protein = True
continue
# Actually make the mutation
bump_score = self.MUT.mutate(resid, new_res, orgtype=old_res)
print 'Mutation: %s, bump_score: %s' % (resid + new_res,
str(bump_score))
if bump_score is None:
skip_protein = True
break
else:
atom_changes = atom_changes + self.MUT.mutate_operations
self.MUT.mutate_operations = []
# Update progress
completion = float(count) / float(numrecords) * 100.0
if callback != None:
callback(completion)
else:
print '%4d of %4d, completion; %5.2f%%' % (
count, float(numrecords), completion)
count = count + 1
# Did it work?
if skip_protein:
print
print 'Modelling failed for %s' % protein
failed.append(protein)
rec.Structure = 'Bumps'
rec.structuretype = 'failed model'
continue
# We have all sets of changes in atom_changes
rec.Structure = {'Rotamer_operations': atom_changes}
rec.Structure['parent'] = parent
rec.structuretype = 'peat model'
print 'Done'
if len(failed) > 0:
print 'Failed to model the following proteins:'
for f in failed:
print f
return
def makemutantSequence(self, sequence, operations):
"""Apply the specified mutations to a sequence and return the mutant seq
Sequence must be in the [[A:0001:ALA],[A:0002:GLU]] format
Operations is a list of the following types:
Mutations: A:0001:ALA:ASP
Deletions: delete:A:0002:GLU
Insertions: insert:1:A:0003:THR:ALA, insert:2:A:0003:THR:TRP (insert THR,TRP after A:0003:THR)
Operations are always performed in sequence numbering order """
if operations == []:
return sequence
ops_sorted = {}
insertions = []
for operation in operations:
s_op = operation.split(':')
# Normal mutation
import pKa.pKD_tools as pKD_tools
resid = pKD_tools.get_resid_from_mut(operation)
if ops_sorted.has_key(resid):
raise Exception(
'More than one operation on the same residue: %s' % resid)
ops_sorted[resid] = ['mutate', operation]
# Perform the operations
new_seq = []
new_count = None
new_chain = None
for resid, restyp in sequence:
# Make sure that the chain hasn't changed or if we are at the beginning then init
if resid.split(':')[0] != new_chain:
#Initialise
sp_resid = resid.split(':')
new_chain = sp_resid[0]
new_count = int(sp_resid[1])
newresid = '%s:%s' % (new_chain, string.zfill(new_count, 4))
# Does this residue have an operation?
if ops_sorted.has_key(resid):
op = ops_sorted[resid]
if op[0] == 'delete':
# Deletion
if op[1] == restyp:
pass # This deletes the residue
else:
raise Exception(
'Incorrect org residue in deletion: %s' % op)
elif op[0] == 'insert':
# Insertion
inserts = op[1].keys()
inserts.sort()
for i in inserts:
if i[0] == restyp:
new_seq.append([newresid, i[1]])
new_count = new_count + 1
newresid = '%s:%s' % (new_chain,
string.zfill(new_count, 4))
elif op[0] == 'mutate':
# Mutation
import pKa.pKD_tools as pKD_tools
orgres = pKD_tools.get_oldrestyp_from_mut(op[1])
if orgres == restyp:
new_seq.append([
newresid,
pKD_tools.get_newrestyp_from_mut(op[1])
])
new_count = new_count + 1
newresid = '%s:%s' % (new_chain,
string.zfill(new_count, 4))
pass
else:
raise Exception('Unknown mutations spec: %s' % op)
else:
new_seq.append([resid, restyp])
new_count = new_count + 1
newresid = '%s:%s' % (new_chain, string.zfill(new_count, 4))
return new_seq
def Model_Mutations(pdbfile,mol2files,mutations,max_overlap=0.5,max_totalbump=1.0,return_score=False,store_mutation_operations=False):
"""Model a number of mutations in a pdbfile when one or more ligands are present"""
#
# Check for stupidity
#
if max_overlap>max_totalbump:
max_totalbump=max_overlap
print 'Adjusted total bump cutoff to %5.2f' %max_totalbump
#
# Read PDB file
#
import Protool
P=Protool.structureIO()
P.readpdb(pdbfile)
P.remove_all_hydrogens()
#
# Read mol2 file
#
L=Protool.ligand(P)
for mol2file in mol2files:
print "Added %s with tag 'LIGAND'" %mol2file
L.readmol2(mol2file,tag='LIGAND')
#
# Get the pdb lines
#
pdblines=P.writepdb('junk.pdb',nowrite=True)
#
# Pass the lines to FFF
#
import FFF.FFFcontrol
myFFF=FFF.FFFcontrol.FFF()
myFFF.parse_lines(pdblines)
#myFFF.soup_stat()
Model=FFF.FFFcontrol.model_class(myFFF,Rotamerlib,FFFaadef_dir)
#
# Store the wild type PDB file
#
if store_mutation_operations:
wt_lines=myFFF.make_pdblines('PDB')
#
import pKa.pKD_tools as pKD_tools
total_bump=0.0
for mutation in mutations:
resid=pKD_tools.get_resid_from_mut(mutation)
chainid=resid.split(':')[0]
resid=resid.split(':')[1]
#
# Get rid of the leading zeros
#
done=False
while not done:
if resid[0]=='0' and len(resid)>1:
resid=resid[1:]
else:
done=True
#
newres=pKD_tools.get_newrestyp_from_mut(mutation)
oldres=pKD_tools.get_oldrestyp_from_mut(mutation)
opttype=3 # Rotamer library
energies=Model.Mutate(chainid,resid,newres,opttype,max_overlap)
bump_score=energies[0]
Sum=energies[1]
Coulomb=energies[2]
#
total_bump=total_bump+bump_score
print 'Bump score: %5.2f, total bump: %5.2f' %(bump_score,total_bump)
if bump_score>max_overlap or total_bump>max_totalbump:
print 'Cannot model this set of mutations - too many bumps'
if return_score:
return False, total_bump
else:
return False
print 'Bump score for %s: %5.3f' %(mutation,bump_score)
print 'Total bump score for all mutations: %5.3f' %(total_bump)
class FFF_fix:
def __init__(self,FFF):
self.PI=FFF
return
#
# Create the instance
#
FFF_instance=FFF_fix(myFFF)
#
# Keep track of the changes that were made to the PDB file
#
if store_mutation_operations:
mut_lines=FFF_instance.PI.make_pdblines('PDB')
import Protool
WT=Protool.structureIO()
WT.parsepdb(wt_lines)
wt_atoms=sorted(WT.atoms.keys())
#
MUT=Protool.structureIO()
MUT.parsepdb(mut_lines)
#
mut_atoms=sorted(MUT.atoms.keys())
wt_count=0
mutcount=0
def coord_diff(atom1,atom2):
diff=0.0
for coord in ['X','Y','Z']:
diff=diff+abs(atom1[coord]-atom2[coord])
return diff
operations=[]
for atom in wt_atoms:
if not atom in mut_atoms:
operations.append(['delete',atom,WT.atoms[atom]])
elif coord_diff(WT.atoms[atom],MUT.atoms[atom])>0.1:
operations.append(['delete',atom,WT.atoms[atom]])
operations.append(['add',atom,MUT.atoms[atom]])
else:
pass
for atom in mut_atoms:
if not atom in wt_atoms:
operations.append(['add',atom,MUT.atoms[atom]])
#
# Store these in FFF_fix
#
FFF_instance.mutate_operations=operations[:]
#
# Return the info
#
if (return_score):
return FFF_instance,total_bump
return FFF_instance