def __add3(self, n_members, excluded, trippleIndex):
"""
Add a tripple of numbers from range( n_members ) to be
excluded for error estimation. Tripples are chosen to have
minimal overlap. For 10 trajectories (*n_members*=10), the
first 3 tripples will be (1,2,3), (4,5,6), (7,8,9).
:param n_members: number of member trajectories
:type n_members: int
:param excluded: excluded member trajectories
:type excluded: [ int ]
:param trippleIndex:
:type trippleIndex: int
:return: the indices of all excluded member trajectories
:rtype: [ int ]
"""
remaining = MU.difference(range(n_members), excluded)
tripple = self.tripples(remaining, trippleIndex + 1)[-1]
return MU.union(excluded, list(tripple))
def __add3( self, n_members, excluded, trippleIndex ):
"""
Add a tripple of numbers from range( n_members ) to be
excluded for error estimation. Tripples are chosen to have
minimal overlap. For 10 trajectories (*n_members*=10), the
first 3 tripples will be (1,2,3), (4,5,6), (7,8,9).
:param n_members: number of member trajectories
:type n_members: int
:param excluded: excluded member trajectories
:type excluded: [ int ]
:param trippleIndex:
:type trippleIndex: int
:return: the indices of all excluded member trajectories
:rtype: [ int ]
"""
remaining = MU.difference( range( n_members ), excluded )
tripple = self.tripples( remaining, trippleIndex+1 )[-1]
return MU.union( excluded, list(tripple) )
def parmSolvated( self, f_out, f_out_crd=None, f_out_pdb=None,
hetatm=0, norun=0,
cap=0, capN=[], capC=[],
fmod=['frcmod.ionsjc_tip3p'], fprep=[],
box=10.0, center=True, **kw ):
"""
:param f_out: target file for parm (topology)
:type f_out: str
:param f_out_crd: target file for crd (coordinates)
(default:|f_out_base|.crd)
:type f_out_crd: str
:param f_out_pdb: target file for pdb (default:|f_out_base|.pdb)
:type f_out_pdb: str
:param hetatm: keep hetero atoms (default: 0)
:type hetatm: 1|0
:param cap: put ACE and NME capping residue on chain breaks
(default: 0)
:type cap: 1|0
:param capN: indices of chains that should get ACE cap (default: [])
:type capN: [int]
:param capC: indices of chains that should get NME cap (default: [])
:type capC: [int]
:param box: minimal distance of solute from box edge (default: 10.0)
:type box: float
:param center: re-center coordinates (default: True)
:type center: bool
:param fmod: list of files with amber parameter modifications
to be loaded into leap with loadAmberParams
(default:['frcmod.ionsjc_tip3p'] ... mod file needed for
default Amber ff10 ions -- topology saving will fail if this
one is missing)
:type fmod: [str]
:param fprep: list of files with amber residue definitions
(to be loaded into leap with loadAmberPrep) (default: [])
:type fprep: [str]
:param kw: additional key=value pairs for leap input template
:type kw: key=value
:raise IOError:
"""
f_out = t.absfile( f_out )
f_out_crd = t.absfile( f_out_crd ) or t.stripSuffix( f_out ) + '.crd'
f_out_pdb = t.absfile( f_out_pdb ) or t.stripSuffix( f_out ) +\
'_leap.pdb'
## removed: (bugfix 3434136)
#fmod = [ t.absfile( f ) for f in t.toList( fmod ) ]
#fprep = [ t.absfile( f ) for f in t.toList( fprep ) ]
try:
if self.verbose: self.log.add( '\nCleaning PDB file for Amber:' )
m = self.leapModel( hetatm=hetatm, center=center )
if cap:
end_broken = m.atom2chainIndices( m.chainBreaks() )
capC = MU.union( capC, end_broken )
capN = MU.union( capN, N0.array( end_broken ) + 1 )
for i in capN:
if self.verbose:
self.log.add( 'Adding ACE cap to chain %i' % i )
m = self.capACE( m, i )
for i in capC:
if self.verbose:
self.log.add( 'Adding NME cap to chain %i' % i )
m = self.capNME( m, i )
m.renumberResidues( addChainId=1 ) ## again, to accomodate capping
template = open( self.leap_template ).read()
leap_mod = self.__fLines( 'm = loadAmberParams %s\n', fmod )
leap_prep= self.__fLines( 'loadAmberPrep %s\n', fprep )
ss = self.__ssBonds( m, cutoff=4. )
self.__cys2cyx( m, ss )
leap_ss = self.__fLines( self.ss_bond, ss )
if self.verbose:
self.log.add('Found %i disulfide bonds: %s' % (len(ss),str(ss)))
if self.verbose:
self.log.add( 'writing cleaned PDB to %s' % self.leap_pdb )
m.writePdb( self.leap_pdb, ter=3 )
self.__runLeap( template, in_pdb=self.leap_pdb,
out_parm=f_out, out_crd=f_out_crd,
ss_bonds=leap_ss, fmod=leap_mod,
fprep=leap_prep, norun=norun,
box=box, **kw )
if not norun:
parm_pdb = self.parm2pdb( f_out, f_out_crd, f_out_pdb )
if not self.keep_leap_pdb and not self.debug:
t.tryRemove( self.leap_pdb )
except IOError as why:
raise IOError(why)
def concat(self, *profiles):
"""
Concatenate all profiles in this with corresponding profiles in the
given ProfileCollection(s). Profiles that are not found in all
ProfileCollections are skipped::
p0.concat( p1 [, p2, ..]) -> single ProfileCollection with the
same number of profiles as p0 but with the length of p0+p1+p2..
:param profiles: profile(s) to concatenate
:type profiles: ProfileCollection(s)
:return: concatenated profile(s)
:rtype: ProfileCollection / subclass
"""
## end recursion (no more arguments)
if len(profiles) == 0:
return self
next = profiles[0]
r = self.__class__()
##!!! BIG FAT WARNING: empty profilecollection does not imply empty model
## an empty PC w/o any profiles currently doesn't know which length
## is is supposed to have. If profLength == 0 for real, then
## the next PC's profiles don't need to be skipped
## Otherwise,
## this creates too-short profiles if the PC parent model has
## non-zero length and simply doesn't have any profiles registered.
## ## special case 1: concat something to empty profile collection
## if not self.keys():
## return next.clone().concat( *profiles[1:] )
##
## ## special case 2: concat empty profile collection to this one
## if not next.keys():
## return self.clone().concat( *profiles[1:] )
##
allkeys = M.union(list(self.profiles.keys()), list(next.keys()))
## for k, p in self.profiles.items():
for k in allkeys:
p = self.profiles.get(k, None)
pnext = next.profiles.get(k, None)
infos = {}
if p is None:
default = next[k, 'default']
p = self.__clonedefault(pnext, self.profLength(), default)
infos = next.infos[k]
if pnext is None:
default = self[k, 'default']
pnext = self.__clonedefault(p, next.profLength(), default)
infos = self.infos[k]
try:
if isinstance(p, N.ndarray):
if len(pnext) == 0:
pnext = pnext.astype(p.dtype)
r.set(k, N.concatenate((p, pnext)), **infos)
else:
r.set(k, p + pnext, **infos)
except:
EHandler.warning("Profile %s skipped during concat." % k,
error=1)
r.remove(k)
return r.concat(*profiles[1:])
def capTerminals( self, auto=False, breaks=False, capN=[], capC=[],
checkgap=True):
"""
Add NME and ACE capping residues to chain breaks or normal N- and
C-terminals. Note: these capping residues contain hydrogen atoms.
Chain indices for capN and capC arguments can be interpreted either
with or without chain break detection enabled. For example, let's
assume we have a two-chain protein with some missing residues (chain
break) in the first chain:
A: MGSKVSK---FLNAGSK
B: FGHLAKSDAK
Then:
capTerminals( breaks=False, capN=[1], capC=[1]) will add N-and
C-terminal caps to chain B.
However:
capTerminals( breaks=True, capN=[1], capC=[1]) will add N- and
C-terminal caps to the second fragment of chain A.
Note: this operation *replaces* the internal model.
:param auto: put ACE and NME capping residue on chain breaks
and on suspected false N- and C-termini (default: False)
:type auto: bool
:param breaks: switch on chain break detection before interpreting
capN and capC
:type breaks: False
:param capN: indices of chains that should get ACE cap (default: [])
:type capN: [int]
:param capC: indices of chains that should get NME cap (default: [])
:type capC: [int]
"""
m = self.model
c_len = m.lenChains()
i_breaks = m.chainBreaks()
if auto:
if not breaks:
capN = self.convertChainIdsNter( m, capN )
capC = self.convertChainIdsCter( m, capC )
breaks=True
capN, capC = self.unresolvedTerminals( m )
end_broken = m.atom2chainIndices( m.chainBreaks(), breaks=1 )
capC = M.union( capC, end_broken )
capN = M.union( capN, N0.array( end_broken ) + 1 )
capN = self.filterProteinChains(m, capN, m.chainIndex(breaks=breaks))
capC = self.filterProteinChains(m, capC, m.chainEndIndex(breaks=breaks))
for i in capN:
m = self.capACE( m, i, breaks=breaks, checkgap=checkgap )
assert m.lenChains() == c_len, '%i != %i' % \
(m.lenChains(), c_len)
assert len(m.chainBreaks(force=True)) == len(i_breaks)
assert m['serial_number'].dtype == N0.Int32, 'serial_number not int'
for i in capC:
m = self.capNME( m, i, breaks=breaks, checkgap=checkgap )
assert m.lenChains() == c_len
assert len(m.chainBreaks(force=True)) == len(i_breaks)
self.model = m
return self.model
def concat( self, *profiles ):
"""
Concatenate all profiles in this with corresponding profiles in the
given ProfileCollection(s). Profiles that are not found in all
ProfileCollections are skipped::
p0.concat( p1 [, p2, ..]) -> single ProfileCollection with the
same number of profiles as p0 but with the length of p0+p1+p2..
:param profiles: profile(s) to concatenate
:type profiles: ProfileCollection(s)
:return: concatenated profile(s)
:rtype: ProfileCollection / subclass
"""
## end recursion (no more arguments)
if len( profiles ) == 0:
return self
next = profiles[0]
r = self.__class__()
##!!! BIG FAT WARNING: empty profilecollection does not imply empty model
## an empty PC w/o any profiles currently doesn't know which length
## is is supposed to have. If profLength == 0 for real, then
## the next PC's profiles don't need to be skipped
## Otherwise,
## this creates too-short profiles if the PC parent model has
## non-zero length and simply doesn't have any profiles registered.
## ## special case 1: concat something to empty profile collection
## if not self.keys():
## return next.clone().concat( *profiles[1:] )
##
## ## special case 2: concat empty profile collection to this one
## if not next.keys():
## return self.clone().concat( *profiles[1:] )
##
allkeys = M.union( list(self.profiles.keys()), list(next.keys()) )
## for k, p in self.profiles.items():
for k in allkeys:
p = self.profiles.get(k, None)
pnext = next.profiles.get(k, None)
infos = {}
if p is None:
default = next[k,'default']
p = self.__clonedefault(pnext, self.profLength(), default)
infos = next.infos[k]
if pnext is None:
default = self[k,'default']
pnext = self.__clonedefault(p, next.profLength(), default)
infos = self.infos[k]
try:
if isinstance( p, N.ndarray ):
if len(pnext) == 0:
pnext = pnext.astype(p.dtype)
r.set( k, N.concatenate( (p, pnext) ), **infos )
else:
r.set( k, p + pnext, **infos )
except:
EHandler.warning("Profile %s skipped during concat." % k,
error=1)
r.remove( k )
return r.concat( *profiles[1:] )
def update(self,
model,
source,
skipRes=None,
updateMissing=0,
force=0,
headPatterns=[]):
"""
Update empty or missing fields of model from the source. The
model will be connected to the source via model.source.
Profiles that are derived from the source are labeled 'changed'=0.
The same holds for coordinates (xyzChanged=0).
However, existing profiles or coordinates or fields remain untouched.
:param model: existing model
:type model: PDBModel
:param source: source PDB file
:type source: str
:param skipRes: list residue names that should not be parsed
:type skipRes: [ str ]
:param updateMissing: ignored
:type updateMissing: 1|0
:param headPatterns: [(putIntoKey, regex)] extract given REMARKS
:type headPatterns: [(str, str)]
:raise PDBParserError - if something is wrong with the source file
"""
try:
## atoms and/or coordinates need to be updated from PDB
if force or self.needsUpdate(model):
atoms, xyz, info = self.__collectAll(source, skipRes,
headPatterns)
keys = M.union(list(atoms.keys()), list(self.DEFAULTS.keys()))
for k in keys:
a = model.atoms.get(k, default=0, update=False)
if (a is 0) or (a is None):
dflt = self.DEFAULTS.get(k, None)
model.atoms.set(k, atoms.get(k, dflt), changed=0)
if model.xyz is None:
model.xyz = xyz
model.xyzChanged = 0
model._resIndex = None
model._chainIndex = None
model.fileName = model.fileName or source
model.pdbCode = model.pdbCode or info.get('pdb_code', None) or \
self.idFromName( model.fileName)
## ## make biounit from the dictionary we have parsed if 'BIOMT' in info:
## biomt = info['BIOMT']
## model.biounit = BU.BioUnit(model, biomt)
## del info['BIOMT']
model.info.update(info)
except:
msg = self.__xplorAtomIndicesTest(source) or ' '
raise PDBParserError('Cannot read ' + str(source) + ' as PDB\n'\
'\ERROR: ' + T.lastError() + msg)
model.setSource(source)
def capTerminals(self,
auto=False,
breaks=False,
capN=[],
capC=[],
checkgap=True):
"""
Add NME and ACE capping residues to chain breaks or normal N- and
C-terminals. Note: these capping residues contain hydrogen atoms.
Chain indices for capN and capC arguments can be interpreted either
with or without chain break detection enabled. For example, let's
assume we have a two-chain protein with some missing residues (chain
break) in the first chain:
A: MGSKVSK---FLNAGSK
B: FGHLAKSDAK
Then:
capTerminals( breaks=False, capN=[1], capC=[1]) will add N-and
C-terminal caps to chain B.
However:
capTerminals( breaks=True, capN=[1], capC=[1]) will add N- and
C-terminal caps to the second fragment of chain A.
Note: this operation *replaces* the internal model.
:param auto: put ACE and NME capping residue on chain breaks
and on suspected false N- and C-termini (default: False)
:type auto: bool
:param breaks: switch on chain break detection before interpreting
capN and capC
:type breaks: False
:param capN: indices of chains that should get ACE cap (default: [])
:type capN: [int]
:param capC: indices of chains that should get NME cap (default: [])
:type capC: [int]
"""
m = self.model
c_len = m.lenChains()
i_breaks = m.chainBreaks()
if auto:
if not breaks:
capN = self.convertChainIdsNter(m, capN)
capC = self.convertChainIdsCter(m, capC)
breaks = True
capN, capC = self.unresolvedTerminals(m)
end_broken = m.atom2chainIndices(m.chainBreaks(), breaks=1)
capC = M.union(capC, end_broken)
capN = M.union(capN, N0.array(end_broken) + 1)
capN = self.filterProteinChains(m, capN, m.chainIndex(breaks=breaks))
capC = self.filterProteinChains(m, capC,
m.chainEndIndex(breaks=breaks))
for i in capN:
m = self.capACE(m, i, breaks=breaks, checkgap=checkgap)
assert m.lenChains() == c_len, '%i != %i' % \
(m.lenChains(), c_len)
assert len(m.chainBreaks(force=True)) == len(i_breaks)
assert m[
'serial_number'].dtype == N0.Int32, 'serial_number not int'
for i in capC:
m = self.capNME(m, i, breaks=breaks, checkgap=checkgap)
assert m.lenChains() == c_len
assert len(m.chainBreaks(force=True)) == len(i_breaks)
self.model = m
return self.model
def parmSolvated( self, f_out, f_out_crd=None, f_out_pdb=None,
hetatm=0, norun=0,
cap=0, capN=[], capC=[],
fmod=['frcmod.ionsjc_tip3p'], fprep=[],
box=10.0, center=True, **kw ):
"""
:param f_out: target file for parm (topology)
:type f_out: str
:param f_out_crd: target file for crd (coordinates)
(default:|f_out_base|.crd)
:type f_out_crd: str
:param f_out_pdb: target file for pdb (default:|f_out_base|.pdb)
:type f_out_pdb: str
:param hetatm: keep hetero atoms (default: 0)
:type hetatm: 1|0
:param cap: put ACE and NME capping residue on chain breaks
(default: 0)
:type cap: 1|0
:param capN: indices of chains that should get ACE cap (default: [])
:type capN: [int]
:param capC: indices of chains that should get NME cap (default: [])
:type capC: [int]
:param box: minimal distance of solute from box edge (default: 10.0)
:type box: float
:param center: re-center coordinates (default: True)
:type center: bool
:param fmod: list of files with amber parameter modifications
to be loaded into leap with loadAmberParams
(default:['frcmod.ionsjc_tip3p'] ... mod file needed for
default Amber ff10 ions -- topology saving will fail if this
one is missing)
:type fmod: [str]
:param fprep: list of files with amber residue definitions
(to be loaded into leap with loadAmberPrep) (default: [])
:type fprep: [str]
:param kw: additional key=value pairs for leap input template
:type kw: key=value
:raise IOError:
"""
f_out = t.absfile( f_out )
f_out_crd = t.absfile( f_out_crd ) or t.stripSuffix( f_out ) + '.crd'
f_out_pdb = t.absfile( f_out_pdb ) or t.stripSuffix( f_out ) +\
'_leap.pdb'
## removed: (bugfix 3434136)
#fmod = [ t.absfile( f ) for f in t.toList( fmod ) ]
#fprep = [ t.absfile( f ) for f in t.toList( fprep ) ]
try:
if self.verbose: self.log.add( '\nCleaning PDB file for Amber:' )
m = self.leapModel( hetatm=hetatm, center=center )
if cap:
end_broken = m.atom2chainIndices( m.chainBreaks() )
capC = MU.union( capC, end_broken )
capN = MU.union( capN, N0.array( end_broken ) + 1 )
for i in capN:
if self.verbose:
self.log.add( 'Adding ACE cap to chain %i' % i )
m = self.capACE( m, i )
for i in capC:
if self.verbose:
self.log.add( 'Adding NME cap to chain %i' % i )
m = self.capNME( m, i )
m.renumberResidues( addChainId=1 ) ## again, to accomodate capping
template = open( self.leap_template ).read()
leap_mod = self.__fLines( 'm = loadAmberParams %s\n', fmod )
leap_prep= self.__fLines( 'loadAmberPrep %s\n', fprep )
ss = self.__ssBonds( m, cutoff=4. )
self.__cys2cyx( m, ss )
leap_ss = self.__fLines( self.ss_bond, ss )
if self.verbose:
self.log.add('Found %i disulfide bonds: %s' % (len(ss),str(ss)))
if self.verbose:
self.log.add( 'writing cleaned PDB to %s' % self.leap_pdb )
m.writePdb( self.leap_pdb, ter=3 )
self.__runLeap( template, in_pdb=self.leap_pdb,
out_parm=f_out, out_crd=f_out_crd,
ss_bonds=leap_ss, fmod=leap_mod,
fprep=leap_prep, norun=norun,
box=box, **kw )
if not norun:
parm_pdb = self.parm2pdb( f_out, f_out_crd, f_out_pdb )
if not self.keep_leap_pdb and not self.debug:
t.tryRemove( self.leap_pdb )
except IOError as why:
raise IOError(why)
def update( self, model, source, skipRes=None, updateMissing=0, force=0,
headPatterns=[]):
"""
Update empty or missing fields of model from the source. The
model will be connected to the source via model.source.
Profiles that are derived from the source are labeled 'changed'=0.
The same holds for coordinates (xyzChanged=0).
However, existing profiles or coordinates or fields remain untouched.
:param model: existing model
:type model: PDBModel
:param source: source PDB file
:type source: str
:param skipRes: list residue names that should not be parsed
:type skipRes: [ str ]
:param updateMissing: ignored
:type updateMissing: 1|0
:param headPatterns: [(putIntoKey, regex)] extract given REMARKS
:type headPatterns: [(str, str)]
:raise PDBParserError - if something is wrong with the source file
"""
try:
## atoms and/or coordinates need to be updated from PDB
if force or self.needsUpdate( model ):
atoms, xyz, info = self.__collectAll( source, skipRes,
headPatterns )
keys = M.union( list(atoms.keys()), list(self.DEFAULTS.keys()) )
for k in keys:
a = model.atoms.get( k, default=0, update=False )
if (a is 0) or (a is None):
dflt = self.DEFAULTS.get( k, None )
model.atoms.set(k, atoms.get(k, dflt), changed=0 )
if model.xyz is None:
model.xyz = xyz
model.xyzChanged = 0
model._resIndex =None
model._chainIndex=None
model.fileName = model.fileName or source
model.pdbCode = model.pdbCode or info.get('pdb_code', None) or \
self.idFromName( model.fileName)
## ## make biounit from the dictionary we have parsed if 'BIOMT' in info:
## biomt = info['BIOMT']
## model.biounit = BU.BioUnit(model, biomt)
## del info['BIOMT']
model.info.update( info )
except:
msg = self.__xplorAtomIndicesTest( source ) or ' '
raise PDBParserError('Cannot read ' + str(source) + ' as PDB\n'\
'\ERROR: ' + T.lastError() + msg)
model.setSource( source )