def checkListStatus( cl, update=0, force_keys=[], version=-1 ):
"""
If the list is to be only updated then check that the list/sublist has
entries that can be updated. If an unupdatable list is sent to
the contactMaster the calculation will halt.
update - 1||0, are we in update mode? if not return whole list
force_keys - 1||0, calculation is restricted to certain keys
"""
if not update:
return cl
## all complexes where force_keys are missing or where values are None
if version == -1:
todo = [ c for c in cl
if None in c.values(keys=force_keys, default=None) ]
else:
todo = [ c[version] for c in cl
if None in c.values(keys=force_keys, default=None) ]
## add missing force keys with value None
if force_keys:
for c in todo:
missing = [ k for k in force_keys if k not in c ]
for k in missing:
c[ k ] = None
if len( todo ) > 0:
msg = "\n%i out of %i complexes are to be updated\n\n"
t.flushPrint( msg%(len(todo), len(cl) ) )
return cl
else:
t.flushPrint( "\nList contains no data that can be updated\n" )
def checkListStatus( cl, update=0, force_keys=[], version=-1 ):
"""
If the list is to be only updated then check that the list/sublist has
entries that can be updated. If an unupdatable list is sent to
the contactMaster the calculation will halt.
update - 1||0, are we in update mode? if not return whole list
force_keys - 1||0, calculation is restricted to certain keys
"""
if not update:
return cl
## all complexes where force_keys are missing or where values are None
if version == -1:
todo = [ c for c in cl
if None in c.values(keys=force_keys, default=None) ]
else:
todo = [ c[version] for c in cl
if None in c.values(keys=force_keys, default=None) ]
## add missing force keys with value None
if force_keys:
for c in todo:
missing = [ k for k in force_keys if k not in c ]
for k in missing:
c[ k ] = None
if len( todo ) > 0:
msg = "\n%i out of %i complexes are to be updated\n\n"
t.flushPrint( msg%(len(todo), len(cl) ) )
return cl
else:
t.flushPrint( "\nList contains no data that can be updated\n" )
def go(self, dict):
"""
Run Modeller job.
@param dict: dictionary with run parameters
@type dict: {param:value}
"""
d = {}
val = None
try:
T.flushPrint( self.params['progress_str'] )
for id, val in dict.items():
modeller_log = LogFile( '%s/Modeller.log' %val["outFolder"] )
d[id] = val
m = M( outFolder= val["outFolder"],
fasta_target=val["fastaTarget"], f_pir=val["f_pir"],
template_folder=val["template_folder"],
starting_model=val["starting_model"],
ending_model=val["ending_model"],
log=modeller_log )
m.run()
except Exception, why:
self.reportError( 'ERROR '+str(why), val )
def pairwise_cast( models ):
"""
atom cast all models in list with each other
-> modified list
"""
for i in range( len( models ) ):
for j in range( i+1, len( models) ):
T.flushPrint('.')
m1 = models[i]
m2 = models[j]
eq_res, eq_atm = m1.equals( m2 )
if not (eq_res and eq_atm):
i1, i2 = m1.compareAtoms( m2 )
delta_1 = len( m1 ) - len( i1 )
delta_2 = len( m2 ) - len( i2 )
models[i].keep( i1 )
models[j].keep( i2 )
f1 = T.stripFilename( m1.sourceFile() )
f2 = T.stripFilename( m2.sourceFile() )
print "Removed %i atoms from %s" % (delta_1, f1)
print "Removed %i atoms from %s" % (delta_2, f2)
return models
def go(self, dict):
"""
Run Modeller job.
@param dict: dictionary with run parameters
@type dict: {param:value}
"""
d = {}
val = None
try:
T.flushPrint(self.params['progress_str'])
for id, val in dict.items():
modeller_log = LogFile('%s/Modeller.log' % val["outFolder"])
d[id] = val
m = M(outFolder=val["outFolder"],
fasta_target=val["fastaTarget"],
f_pir=val["f_pir"],
template_folder=val["template_folder"],
starting_model=val["starting_model"],
ending_model=val["ending_model"],
log=modeller_log)
m.run()
except Exception, why:
self.reportError('ERROR ' + str(why), val)
def pairwise_cast(models):
"""
atom cast all models in list with each other
-> modified list
"""
for i in range(len(models)):
for j in range(i + 1, len(models)):
T.flushPrint('.')
m1 = models[i]
m2 = models[j]
eq_res, eq_atm = m1.equals(m2)
if not (eq_res and eq_atm):
i1, i2 = m1.compareAtoms(m2)
delta_1 = len(m1) - len(i1)
delta_2 = len(m2) - len(i2)
models[i].keep(i1)
models[j].keep(i2)
f1 = T.stripFilename(m1.sourceFile())
f2 = T.stripFilename(m2.sourceFile())
print "Removed %i atoms from %s" % (delta_1, f1)
print "Removed %i atoms from %s" % (delta_2, f2)
return models
def go(self, cmplxDic):
"""
Obtain contact matrix for all complexes.
@param cmplxDic: dictionary of complexes::
{soln:Complex, soln:Complex, ...}
@type cmplxDic: {int:Complex}
@return: similar dictionary with Complex.info['soln'] as keys and
file names of matrices as value::
{ soln : fname, soln : fname ....}
@rtype: {int:str}
"""
result = {}
startTime = time.time()
for soln, c in cmplxDic.items():
T.flushPrint("%i," % soln)
## if not os.path.exists( T.absfile('~/debug.dic') ):
## T.dump( cmplxDic, T.absfile('~/debug.dic') )
self.calcContacts(soln, c)
self.calcInterfaceRms(soln, c)
self.calcReducedContacts(soln, c)
## TODO: Prosa will not run when called via conatacSlave, runs as it should when
## called as c.prosa2003Energy() in the interpreter. What's wroong here?
## For mow the Prosa calculation is skipped.
##
self.calcProsa(soln, c)
self.calcPairScore(soln, c) ## uses res-contacts
self.calcFoldX(soln, c) ##uses rec/lig.info['foldX'] if available
for method in ['cons_ent', 'cons_max', 'cons_abs']:
self.calcConservation(soln, c, method)
c['__version_contacter'] = self.version()
result[soln] = c
c.slim()
## if not os.path.exists(T.absfile('~/debug_afterslave.dic') ):
## T.dump( cmplxDic, T.absfile('~/debug_afterslave.dic') )
print "\navg time for last %i complexes: %f s" %\
( len(cmplxDic), (time.time()-startTime)/len(cmplxDic))
return result
def go(self, cmplxDic):
"""
Obtain contact matrix for all complexes.
@param cmplxDic: dictionary of complexes::
{soln:Complex, soln:Complex, ...}
@type cmplxDic: {int:Complex}
@return: similar dictionary with Complex.info['soln'] as keys and
file names of matrices as value::
{ soln : fname, soln : fname ....}
@rtype: {int:str}
"""
result = {}
startTime = time.time()
for soln, c in cmplxDic.items():
T.flushPrint( "%i," % soln )
## if not os.path.exists( T.absfile('~/debug.dic') ):
## T.dump( cmplxDic, T.absfile('~/debug.dic') )
self.calcContacts( soln, c )
self.calcInterfaceRms( soln, c )
self.calcReducedContacts( soln, c )
## TODO: Prosa will not run when called via conatacSlave, runs as it should when
## called as c.prosa2003Energy() in the interpreter. What's wroong here?
## For mow the Prosa calculation is skipped.
##
self.calcProsa( soln, c )
self.calcPairScore( soln, c ) ## uses res-contacts
self.calcFoldX( soln, c ) ##uses rec/lig.info['foldX'] if available
for method in ['cons_ent', 'cons_max', 'cons_abs']:
self.calcConservation( soln, c, method )
c['__version_contacter'] = self.version()
result[ soln ] = c
c.slim()
## if not os.path.exists(T.absfile('~/debug_afterslave.dic') ):
## T.dump( cmplxDic, T.absfile('~/debug_afterslave.dic') )
print "\navg time for last %i complexes: %f s" %\
( len(cmplxDic), (time.time()-startTime)/len(cmplxDic))
return result
def sloppyload( f ):
"""
f - str, file name
-> any, unpickled object
"""
try:
T.flushPrint( "Loading " + str(f) + '\n' )
return T.load( T.absfile( f ) )
except cPickle.UnpicklingError:
print "Trying to load %s in sloppy mode..." % f
return PickleUpgrader(open(T.absfile(f))).load()
def sloppyload( f ):
"""
f - str, file name
-> any, unpickled object
"""
try:
T.flushPrint( "Loading " + str(f) + '\n' )
return T.load( T.absfile( f ) )
except cPickle.UnpicklingError:
print "Trying to load %s in sloppy mode..." % f
return PickleUpgrader(open(T.absfile(f))).load()
def go(self, dict):
"""
Run alignment job.
@param dict: dictionary with run parameters
@type dict: {param:value}
"""
d = {}
val = None
try:
T.flushPrint(self.progress_str)
for id, val in dict.items():
aligner_log = LogFile('%s/Aligner.log' % val["outFolder"])
d[id] = val
aligner_log.add('Slave aligns %s on %s' % (id, os.uname()[1]))
a = Aligner(outFolder=val["outFolder"], log=aligner_log)
## For the cross validation
if not os.path.exists(val["outFolder"] + TC.F_COFFEE):
input_file = val["outFolder"] + VS.F_TCOFFEE
alpha_path = self.prepareT_coffee(input_file)
a.align_for_modeller_inp(
pdbFiles=alpha_path,
fasta_templates=val["fastaTemplates"],
fasta_sequences=val["fastaSequences"],
fasta_target=val["fastaTarget"])
## For a classic project folder
else:
a.align_for_modeller_inp(
pdbFiles=val["pdbFiles"],
fasta_templates=val["fastaTemplates"],
fasta_sequences=val["fastaSequences"],
fasta_target=val["fastaTarget"])
a.go()
except Exception, why:
self.reportError('ERROR ' + str(why), val)
def go(self, dict):
"""
Run alignment job.
@param dict: dictionary with run parameters
@type dict: {param:value}
"""
d = {}
val = None
try:
T.flushPrint( self.progress_str )
for id, val in dict.items():
aligner_log = LogFile( '%s/Aligner.log' %val["outFolder"] )
d[id] = val
aligner_log.add('Slave aligns %s on %s' % (id,os.uname()[1]) )
a = Aligner( outFolder= val["outFolder"], log=aligner_log)
## For the cross validation
if not os.path.exists(val["outFolder"] + TC.F_COFFEE):
input_file = val["outFolder"] + VS.F_TCOFFEE
alpha_path = self.prepareT_coffee(input_file)
a.align_for_modeller_inp( pdbFiles=alpha_path,
fasta_templates=val["fastaTemplates"],
fasta_sequences=val["fastaSequences"],
fasta_target=val["fastaTarget"])
## For a classic project folder
else:
a.align_for_modeller_inp(pdbFiles=val["pdbFiles"],
fasta_templates=val["fastaTemplates"],
fasta_sequences=val["fastaSequences"],
fasta_target=val["fastaTarget"])
a.go()
except Exception, why:
self.reportError( 'ERROR '+str(why), val )
def go(self, dict):
"""
Perform slave task.
"""
d = {}
T.flushPrint(self.params['progress_str'])
for id, val in dict.items():
d[id] = val + 1
for i in range(1, 1000):
f = 1.0 * i / 1200232112312.11
T.flushPrint(str(id) + ' ')
time.sleep(0.5)
print "Done."
return d
def go(self, dict):
"""
Perform slave task.
"""
d = {}
T.flushPrint( self.params['progress_str'] )
for id, val in dict.items():
d[id] = val+1
for i in range(1, 1000):
f = 1.0 * i / 1200232112312.11
T.flushPrint( str(id) + ' ' )
time.sleep(0.5)
print "Done."
return d
def setHexComplexes(self, com_lst, n=20 ):
"""
add contact matrices of hex-generated (wrong) complexes for comparison
@param com_lst: ComplexList with contacts calculated
@type com_lst: ComplexList
"""
t.flushPrint('adding hex-generated complexes')
self.hexContacts = []
# f = lambda a: molUtils.elementType(a['element']) == 'p'
i = 0
while i < n:
com = com_lst[i]
t.flushPrint('#')
try:
if com['fractNatCont'] == 0.0:#com['fnac_10'] == 0.0:
com.rec_model.remove( com.rec().maskH() )
com.lig_model.remove( com.lig_model.maskH() )
com.rec_model = com.rec_model.sort()
com.lig_model = com.lig_model.sort()
com.rec_model.keep( self.i_free_rec )
com.lig_model.keep( self.i_free_lig )
com.lig_transformed = None
self.hexContacts += [ com.atomContacts() ]
else:
n += 1
i+= 1
except:
print t.lastError()
self.hexSurfaces = self.__categorizeHexSurf( 0.2 )
def go(self, jobs):
"""
The calculation.
@param jobs: dictionary with { int_id : str_protocol }
@type jobs: dict
@return: result from AmberEntropist.run()
@rtype: dict
"""
result = {}
startTime = time.time()
for id, protocol in jobs.items():
try:
T.flushPrint("%s " % str(id))
protocol.update({'nice': self.nice})
x = None ## free memory from previous run
x = AmberEntropist(**protocol)
x.run()
r = x.result
if r:
r['__version_AmberEntropist'] = x.version()
result[id] = r
else:
result[id] = None
except EntropistError, why:
self.reportError(str(type(why)), id)
except IOError, why:
self.reportError(str(why), id)
def go(self, jobs):
"""
The calculation.
@param jobs: dictionary with { int_id : str_protocol }
@type jobs: dict
@return: result from AmberEntropist.run()
@rtype: dict
"""
result = {}
startTime = time.time()
for id, protocol in jobs.items():
try:
T.flushPrint( "%s " % str(id) )
protocol.update( {'nice':self.nice} )
x = None ## free memory from previous run
x = AmberEntropist( **protocol )
x.run()
r = x.result
if r:
r['__version_AmberEntropist'] = x.version()
result[ id ] = r
else:
result[ id ] = None
except EntropistError, why:
self.reportError( str(type(why)), id )
except IOError, why:
self.reportError( str(why), id )
def go(self, jobs):
"""
Run job.
@param jobs: { ((int,int),(int,int)) : (str, str) }, maps start and end
position of two chunks of coordinate frames to the
files where the two chunks are pickled.
@type jobs: {((int,int),(int,int)) : (str, str)}
@return: the rms between the frames
@rtype: [float]
"""
result = {}
frames = None
startTime = time.time()
try:
for i, frames in jobs.items():
T.flushPrint( str(i) )
f1 = self.__getFrames( frames[0])
f2 = self.__getFrames( frames[1])
result[ i ] = self.calcRmsd(i, f1, f2 )
print "\navg time for last %i complexes: %f s" %\
( len(jobs), (time.time()-startTime)/len(jobs))
except IOError, why:
self.reportError("Cannot open temporary frame file "+\
"(can happen if slave catches exit signal)",
frames )
def reduceComplexList( cl ):
"""
If cl comes from a multiple model docking, and HEX macrodocking was
switched on, the list contains more than 512 solutions per model
combination. Keep only the first 512 solutions.
-> shortened or unchanged ComplexList
"""
if len( cl ) != len(cl.models.recModels()) * len(cl.models.ligModels()) * 512:
T.flushPrint('\nRemoving HEX solutions greater than 512.\n')
len_old = len( cl )
cl = cl.filter('soln', (0,512) )
T.flushPrint('%i solutions removed.\n' % (len_old - len(cl) ) )
T.flushPrint('%i solutions remain.\n' % len( cl ) )
return cl
def reduceComplexList(cl):
"""
If cl comes from a multiple model docking, and HEX macrodocking was
switched on, the list contains more than 512 solutions per model
combination. Keep only the first 512 solutions.
-> shortened or unchanged ComplexList
"""
if len(cl) != len(cl.models.recModels()) * len(
cl.models.ligModels()) * 512:
T.flushPrint('\nRemoving HEX solutions greater than 512.\n')
len_old = len(cl)
cl = cl.filter('soln', (0, 512))
T.flushPrint('%i solutions removed.\n' % (len_old - len(cl)))
T.flushPrint('%i solutions remain.\n' % len(cl))
return cl
def reduceComplexList(cl):
"""
If cl comes from a multiple model docking, and HEX macrodocking was
switched on, the list contains more than 512 solutions per model
combination. Keep only the first 512 solutions.
-> shortened or unchanged ComplexList
"""
if len(cl) == len(cl.recModels()) * len(cl.ligModels()) * 512:
return cl
if len(cl) < 512:
print '\nNOTE: THE COMPLEX LIST IS SHORTER THAN 512 COMPLEXES'
print ' COMPLEX LIST CONTAINS %i ONLY COMPLEXES\n' % len(cl)
return cl
t.flushPrint('\nRemoving HEX solutions greater than 512.\n')
r = cl.filter('soln', (0, 512))
t.flushPrint('%i solutions removed.\n' % (len(cl) - len(r)))
t.flushPrint('%i solutions remain.\n' % len(r))
return r
def reduceComplexList( cl ):
"""
If cl comes from a multiple model docking, and HEX macrodocking was
switched on, the list contains more than 512 solutions per model
combination. Keep only the first 512 solutions.
-> shortened or unchanged ComplexList
"""
if len( cl ) == len(cl.recModels()) * len(cl.ligModels()) * 512:
return cl
if len( cl ) < 512:
print '\nNOTE: THE COMPLEX LIST IS SHORTER THAN 512 COMPLEXES'
print ' COMPLEX LIST CONTAINS %i ONLY COMPLEXES\n'%len(cl)
return cl
t.flushPrint('\nRemoving HEX solutions greater than 512.\n')
r = cl.filter('soln', (0,512) )
t.flushPrint('%i solutions removed.\n' % (len( cl ) - len( r ) ) )
t.flushPrint('%i solutions remain.\n' % len( r ) )
return r
###########################
# MAIN
###########################
if len(sys.argv) < 3:
_use()
options = t.cmdDict(defaultOptions())
## ## current keys used for scoring
## scoreKeys = ['eProsa', 'ePairScore', 'foldX', 'cons_ent', 'cons_max']
## load docking solutions
t.flushPrint("\nLoading complex list %s ... " % t.absfile(options['i']))
complex_lst = t.load(options['i'])
t.flushPrint("done\n")
## validate and expand list of keys to be calculated
force = []
if options.has_key('f'):
raw_force = t.toList(options['f'])
## check that the key is valid
validKeys = [
'fnac_4.5', 'fnac_10', 'fnrc_4.5', 'fnarc_9', 'fnarc_10', 'c_ratom_9',
'c_ratom_10', 'eProsa', 'ePairScore', 'foldX', 'cons_ent', 'cons_max',
'cons_abs', 'rms_if', 'rms_if_bb'
]
## Find homologues to the target sequence using blast against "seq_db"
## Cluster the seuences and write the result to nr.fasta
## input: target.fasta
##
## output: sequences/all.fasta
## /blast.out
## /cluster_result.out
## /nr.fasta (input for Aligner)
searcher = SequenceSearcher( outFolder=outFolder, clusterLimit=clustLim,
verbose=1, log=log )
if 'psi' in options:
## local PSIBlast - not fully implemented!!
tools.flushPrint('Performing local PSI blast search\n')
rounds = int( options['psi'] )
searcher.localPSIBlast( f_target, seq_db, e=e, rounds=rounds,
alignments=aln, **ext_options )
else:
## if it looks like local Blast is installed
tools.flushPrint('Performing local blast search\n')
if os.environ.has_key('BLASTDB') and settings.blast_bin:
searcher.localBlast( f_target, seq_db, 'blastp', alignments=aln,
e=e, **ext_options )
## try remote Blast
else:
tools.flushPrint('Performing remote blast search\n')
searcher.remoteBlast( f_target, seq_db, 'blastp', alignments=aln,
e=e, **ext_options )
pdb - PDB code to be stored in trajectory
"""
sys.exit(0)
##########
## MAIN ##
use()
o = T.cmdDict( {'n':10} )
f_in = T.absfile( o['i'] )
f_out = T.absfile( o.get('o', f_in) )
n = int( o['n'] )
T.flushPrint("Loading...")
t = T.load( f_in )
T.flushPrint("Converting %i frames..." % len(t) )
if isinstance(t, EnsembleTraj ):
T.flushPrint( "Nothing to be done!\n")
sys.exit(0)
t = traj2ensemble( t, n )
if 'pdb' in o:
t.ref.pdbCode = o['pdb']
if f_in == f_out:
os.rename( f_in, f_in + '_backup')
print 'Current directory is not a valid modeling project folder.'
_use( options )
## Try to add project folders
## look for default cross-validation projects
d = []
if os.path.exists( f + VS.F_RESULT_FOLDER ):
d = [ f ]
if options.has_key('d'):
folders = T.toList(options['d'])
else:
folders = d
T.flushPrint("Starting job...\n")
for f in folders:
a = A(outFolder=f)
a.go()
T.flushPrint("Done.\n")
## show result in PyMol
if options.has_key('s'):
p=Pymoler()
p.addPdb( folders[0] + a.F_FINAL_PDB )
p.add('color_b')
p.add('select na, b<0')
p.add('color grey, na')
p.add('hide all')
'rdic': 'pcr_rec/models.dic'
})
if len(sys.argv) < 2:
syntax()
fs = [T.absfile(f) for f in T.toList(options['i'])]
result = ComplexList()
rec_dic = T.load(T.absfile(options['rdic']))
lig_dic = T.load(T.absfile(options['ldic']))
for f in fs:
T.flushPrint('Loading %s ...' % f)
cl = T.load(f)
cl = reduceComplexList(cl)
result += cl
T.flushPrint('done\n')
T.flushPrint('correct model numbers...')
correct_model_numbers(result, rec_dic, lig_dic)
T.flushPrint('\ncasting all rec models...')
pairwise_cast(result.models.recModels())
Default options:
"""
for key, value in o.items():
print "\t-",key, "\t",value
sys.exit(0)
if __name__ == '__main__':
options = T.cmdDict({'o':[ os.getcwd() ]})
if '?' in options or 'help' in options:
_use( options )
folders = T.toList( options['o'] )
if not os.path.exists( folders[0] +'/templates'):
print 'Current directory is not a valid modeling folder.'
_use( options )
T.flushPrint( "Creating folders and links...\n" )
for f in folders:
sv = VS(outFolder=f)
sv.go(f)
T.flushPrint( "done\n" )
###############
## SequenceSearcher
##
## Find homologues to the target sequence using blast against "seq_db"
## Cluster the seuences and write the result to nr.fasta
## input: target.fasta
##
## output: sequences/all.fasta
## /blast.out
## /cluster_result.out
## /nr.fasta (input for Aligner)
tools.flushPrint('Searching for homologues ...')
try:
# initiate
searcher = SequenceSearcher( outFolder=outFolder, verbose=1, log=log )
# ## local PSIBlast - not fully implemented!!
# searcher.localPSIBlast( target, seq_db, e=0.1, alignments=1000)
## local Blast
searcher.localBlast( f_target, seq_db, 'blastp', alignments=500, e=0.0001 )
## cluster blast results. Defaults: simCut=1.75, lenCut=0.9, ncpu=1
## expects all.fasta
# searcher.clusterFastaIterative( )
searcher.clusterFasta()
if not os.path.exists(f + VS.F_RESULT_FOLDER) and not options.has_key('d'):
print 'Current directory is not a valid modeling project folder.'
_use(options)
## Try to add project folders
## look for default cross-validation projects
d = []
if os.path.exists(f + VS.F_RESULT_FOLDER):
d = [f]
if options.has_key('d'):
folders = T.toList(options['d'])
else:
folders = d
T.flushPrint("Starting job...\n")
for f in folders:
a = A(outFolder=f)
a.go()
T.flushPrint("Done.\n")
## show result in PyMol
if options.has_key('s'):
p = Pymoler()
p.addPdb(folders[0] + a.F_FINAL_PDB)
p.add('color_b')
p.add('select na, b<0')
p.add('color grey, na')
p.add('hide all')
_use(options)
if not os.path.exists(f + B.F_INPUT_FOLDER) and not options.has_key('d'):
print 'Current directory is not a valid modeling project folder.'
_use(options)
## Try to add project folders
## look for default cross-validation projects
d = []
if os.path.exists(f + VS.F_RESULT_FOLDER):
d = glob.glob(f + VS.F_RESULT_FOLDER + '/*')
if options.has_key('d'):
folders = T.toList(options['d'])
else:
folders = d
reference = options.get('ref', None)
model_list = options.get('modlist', None)
T.flushPrint("Starting job...\n")
for f in folders:
T.flushPrint("\tWorking on %s\n" % os.path.split(f)[1])
b = B(outFolder=f)
b.go(model_list=model_list, reference=reference)
T.flushPrint("Done.\n")
if end:
end = int( end )
step = int( o.get('step',1) )
ref = o.get('ref',None)
if ref:
ref = PDBModel( T.absfile( ref ) )
if 'prot' in o:
ref = ref.compress( ref.maskProtein() )
result_xyz = []
result_frameNames = []
result_ref = None
T.flushPrint("Loading and appending trajectories...")
for i in range( len( inLst ) ):
t = loadTraj( inLst[i], i, start, end, step )
if 'prot' in o:
t.keepAtoms( N0.nonzero(t.ref.maskProtein()) )
result_ref = result_ref or ref or t.ref
if t.ref.equals( result_ref ) != [1,1]:
raise Exception( 'Incompatible reference structure.' )
for xyz in t.frames:
result_xyz.append( xyz.astype('f') )
except cPickle.UnpicklingError:
print "Trying to load %s in sloppy mode..." % f
return PickleUpgrader(open(T.absfile(f))).load()
#########
## Main
#########
__use()
fs = sys.argv[1:]
for f in fs:
try:
o = sloppyload( f )
## don't slim PDBModels that are their own source
if isinstance( o, PDBModel ) and str( o.source ) == T.absfile( f ):
o.forcePickle = 1
T.flushPrint('Dumping %s\n' % f )
T.dump( o, T.absfile( f ) )
except:
print "Error with ", f
print T.lastError()
## input: target.fasta
##
## output: sequences/all.fasta
## /blast.out
## /cluster_result.out
## /nr.fasta (input for Aligner)
searcher = SequenceSearcher(outFolder=outFolder,
clusterLimit=clustLim,
verbose=1,
log=log)
if 'psi' in options:
## local PSIBlast - not fully implemented!!
tools.flushPrint('Performing local PSI blast search\n')
rounds = int(options['psi'])
searcher.localPSIBlast(f_target,
seq_db,
e=e,
rounds=rounds,
alignments=aln,
**ext_options)
else:
## if it looks like local Blast is installed
tools.flushPrint('Performing local blast search\n')
if os.environ.has_key('BLASTDB') and settings.blast_bin:
searcher.localBlast(f_target,
seq_db,
'blastp',
alignments=aln,
def retrievePDBs(self, outFolder=None, pdbCodes=None):
"""
Get PDB from local database if it exists, if not try to
download the coordinartes drom the RSCB.
Write PDBs for given fasta records. Add PDB infos to internal
dictionary of fasta records. NMR structures get resolution 3.5.
@param outFolder: folder to put PDB files into (default: L{F_ALL})
@type outFolder: str OR None
@param pdbCodes: list of PDB codes [all previously found templates]
@type pdbCodes: [str]
@return: list of PDB file names
@rtype: [str]
@raise BlastError: if can't write file
"""
outFolder = outFolder or self.outFolder + self.F_ALL
pdbCodes = pdbCodes or self.record_dic.keys()
result = []
i = 0
if not self.silent:
T.flushPrint("fetching %i PDBs (l=local, r=remotely)..." % \
len( pdbCodes ) )
for c in pdbCodes:
i += 1
fname = '%s/%s.pdb' % (outFolder, c)
try:
if os.path.exists(fname):
h = open(fname, 'r')
else:
h = self.getLocalPDBHandle(c)
if not self.silent:
T.flushPrint('l')
except:
h = self.getRemotePDBHandle(c)
if not self.silent:
T.flushPrint('r')
try:
lines, infos = self.parsePdbFromHandle(h, first_model_only=1)
infos['file'] = fname
if c in self.record_dic:
self.record_dic[c].__dict__.update(infos)
## close if it is a handle
try:
h.close()
except:
pass
if not os.path.exists(fname):
f = open(fname, 'w', 1)
f.writelines(lines)
f.close()
result += [fname]
except IOError, why:
raise BlastError("Can't write file " + fname)
"""
sys.exit(0)
##########
## MAIN ##
use()
o = T.cmdDict({'n': 10})
f_in = T.absfile(o['i'])
f_out = T.absfile(o.get('o', f_in))
n = int(o['n'])
T.flushPrint("Loading...")
t = T.load(f_in)
T.flushPrint("Converting %i frames..." % len(t))
if isinstance(t, EnsembleTraj):
T.flushPrint("Nothing to be done!\n")
sys.exit(0)
t = traj2ensemble(t, n)
if 'pdb' in o:
t.ref.pdbCode = o['pdb']
if f_in == f_out:
os.rename(f_in, f_in + '_backup')
Default options:
"""
for key, value in o.items():
print "\t-", key, "\t", value
sys.exit(0)
if __name__ == '__main__':
options = T.cmdDict({'o': [os.getcwd()]})
if '?' in options or 'help' in options:
_use(options)
folders = T.toList(options['o'])
if not os.path.exists(folders[0] + '/templates'):
print 'Current directory is not a valid modeling folder.'
_use(options)
T.flushPrint("Creating folders and links...\n")
for f in folders:
sv = VS(outFolder=f)
sv.go(f)
T.flushPrint("done\n")
except cPickle.UnpicklingError:
print "Trying to load %s in sloppy mode..." % f
return PickleUpgrader(open(T.absfile(f))).load()
#########
## Main
#########
__use()
fs = sys.argv[1:]
for f in fs:
try:
o = sloppyload( f )
## don't slim PDBModels that are their own source
if isinstance( o, PDBModel ) and str( o.source ) == T.absfile( f ):
o.forcePickle = 1
T.flushPrint('Dumping %s\n' % f )
T.dump( o, T.absfile( f ) )
except:
print "Error with ", f
print T.lastError()
"""
Restart a distributed calculation.
Syntax: restartPVM.py -i |rst_file| [-a]
Options:
i .. restart file containing result of TrackingJobMaster.getRst()
a .. add hosts to PVM
"""
sys.exit(0)
## MAIN ##
use()
cmd = T.cmdDict()
T.flushPrint('Loading restart data...')
rst = T.load( cmd['i'] )
hosts = []
if 'a' in cmd:
hosts = [ h['host'] for h in rst['hosts'] ]
master = restart( rst, hosts=hosts )
T.flushPrint('Master initialized for restart.')
master.start()
if not os.path.exists( f + B.F_INPUT_FOLDER ) and not options.has_key('d'):
print 'Current directory is not a valid modeling project folder.'
_use( options )
## Try to add project folders
## look for default cross-validation projects
d = []
if os.path.exists( f + VS.F_RESULT_FOLDER ):
d = glob.glob( f + VS.F_RESULT_FOLDER + '/*' )
if options.has_key('d'):
folders = T.toList(options['d'])
else:
folders = d
reference = options.get('ref', None)
model_list = options.get('modlist', None)
T.flushPrint("Starting job...\n")
for f in folders:
T.flushPrint("\tWorking on %s\n"%os.path.split(f)[1])
b = B( outFolder=f )
b.go(model_list = model_list, reference = reference)
T.flushPrint( "Done.\n")
end = o.get('e', None)
if end:
end = int(end)
step = int(o.get('step', 1))
ref = o.get('ref', None)
if ref:
ref = PDBModel(T.absfile(ref))
if 'prot' in o:
ref = ref.compress(ref.maskProtein())
result_xyz = []
result_frameNames = []
result_ref = None
T.flushPrint("Loading and appending trajectories...")
for i in range(len(inLst)):
t = loadTraj(inLst[i], i, start, end, step)
if 'prot' in o:
t.keepAtoms(N.nonzero(t.ref.maskProtein()))
result_ref = result_ref or ref or t.ref
if t.ref.equals(result_ref) != [1, 1]:
raise Exception('Incompatible reference structure.')
for xyz in t.frames:
result_xyz.append(xyz.astype('f'))
if len(sys.argv) < 2:
print \
"""
Restart a distributed calculation.
Syntax: restartPVM.py -i |rst_file| [-a]
Options:
i .. restart file containing result of TrackingJobMaster.getRst()
a .. add hosts to PVM
"""
sys.exit(0)
## MAIN ##
use()
cmd = T.cmdDict()
T.flushPrint('Loading restart data...')
rst = T.load(cmd['i'])
hosts = []
if 'a' in cmd:
hosts = [h['host'] for h in rst['hosts']]
master = restart(rst, hosts=hosts)
T.flushPrint('Master initialized for restart.')
master.start()
def randomSurfaces( base_folder, label, mask ):
"""
calculate surfaces for all peptides and return the
average and SD
"""
## container for results and standard deviations
MS, AS = {}, {}
MS_sd, AS_sd = {}, {}
## loop over peptide directories
for k in MOU.aaAtoms.keys():
dir = base_folder + 'GLY-%s-GLY_pcr/pcr_00'%(k)
fLst = glob.glob( dir + '/*.pdb')
msLst = []
asLst = []
## loop over pdb files for each peptide
T.flushPrint( '\nNow collecting data in %s'%dir )
for f in fLst:
## load peptide and remove waters and hydrogens
m = PDBModel( f )
m = m.compress( m.maskProtein() * m.maskHeavy() )
T.flushPrint( '.')
## add surface data
try:
d = PDBDope( m )
d.addSurfaceRacer( probe=1.4 )
## remove tailing GLY
m = m.compress( m.res2atomMask(mask) )
## collect surface data for each peptide
msLst += [ m.profile('MS') ]
asLst += [ m.profile('AS') ]
except:
print 'Failed calculating exposure for GLY-%s-GLY'%(k)
print '\t and file %s'%f
## get result dictionary for peptide
T.flushPrint('\nCollecting data ...\n')
msDic = {}
asDic = {}
msDic_sd = {}
asDic_sd = {}
j = 0
#atoms = [ a['name'] for a in m.atoms ]
for n in m['name']:
msDic[n] = N0.average(msLst)[j]
asDic[n] = N0.average(asLst)[j]
msDic_sd[n] = MAU.SD( msLst )[j]
asDic_sd[n] = MAU.SD( asLst )[j]
j += 1
MS[ k ] = msDic
AS[ k ] = asDic
MS_sd[ k ] = msDic_sd
AS_sd[ k ] = asDic_sd
return MS, AS, MS_sd, AS_sd
tmp_db = 'pdbaa'
###############
## SequenceSearcher
##
## Find homologues to the target sequence using blast against "seq_db"
## Cluster the seuences and write the result to nr.fasta
## input: target.fasta
##
## output: sequences/all.fasta
## /blast.out
## /cluster_result.out
## /nr.fasta (input for Aligner)
tools.flushPrint('Searching for homologues ...')
try:
# initiate
searcher = SequenceSearcher(outFolder=outFolder, verbose=1, log=log)
# ## local PSIBlast - not fully implemented!!
# searcher.localPSIBlast( target, seq_db, e=0.1, alignments=1000)
## local Blast
searcher.localBlast(f_target, seq_db, 'blastp', alignments=500, e=0.0001)
## cluster blast results. Defaults: simCut=1.75, lenCut=0.9, ncpu=1
## expects all.fasta
# searcher.clusterFastaIterative( )
searcher.clusterFasta()
def __init__( self, verbose=1, **options ):
"""
@param verbose: verbosity level (default: 1)
@type verbose: 1|0
@param options: needs::
rec,lig - file name, receptor, ligand trajectories
ref - file name, pickled reference complex
@type options: any
@raise AnalyzeError: if atoms are not aligned
"""
self.options = options
if verbose: t.flushPrint("\nLoading...")
self.t_lig = t.load( options['lig'] )
self.t_rec = t.load( options['rec'] )
self.com= t.load( options['ref'] )
## delete H from all players
self.t_lig.removeAtoms( self.t_lig.getRef().maskH() )
self.t_rec.removeAtoms( self.t_rec.getRef().maskH() )
self.com.rec_model.remove( self.com.rec_model.maskH() )
self.com.lig_model.remove( self.com.lig_model.maskH() )
self.com.lig_transformed = None
## equalize atom content of free (trajectory) and bound models
if verbose: t.flushPrint('\nCasting...')
bnd_rec = self.com.rec()
bnd_lig = self.com.lig_model
self.t_rec.sortAtoms()
self.t_lig.sortAtoms()
bnd_rec = bnd_rec.sort()
bnd_lig = bnd_lig.sort()
#m_bnd_rec, m_t_rec = bnd_rec.equalAtoms( self.t_rec.getRef() )
#m_bnd_lig, m_t_lig = bnd_lig.equalAtoms( self.t_lig.getRef() )
i_bnd_rec, i_t_rec = bnd_rec.compareAtoms( self.t_rec.getRef() )
i_bnd_lig, i_t_lig = bnd_lig.compareAtoms( self.t_lig.getRef() )
#self.t_rec.removeAtoms( N0.logical_not( m_t_rec ) )
#self.t_lig.removeAtoms( N0.logical_not( m_t_lig ) )
self.t_rec = self.t_rec.takeAtoms( i_t_rec )
self.t_lig = self.t_lig.takeAtoms( i_t_lig )
#self.mask_free_lig = m_t_lig
#self.mask_free_rec = m_t_rec
self.i_free_lig = i_t_lig
self.i_free_rec = i_t_rec
#bnd_rec.remove( N0.logical_not( m_bnd_rec ) )
#bnd_lig.remove( N0.logical_not( m_bnd_lig ) )
bnd_rec = bnd_rec.take( i_bnd_rec )
bnd_lig = bnd_lig.take( i_bnd_lig )
#self.mask_bnd_rec = m_bnd_rec
#self.mask_bnd_lig = m_bnd_lig
self.i_bnd_rec = i_bnd_rec
self.i_bnd_lig = i_bnd_lig
## put 'equalized' models back into ref complex
self.com.rec_model = bnd_rec
self.com.lig_model = bnd_lig
self.com.lig_transformed = None
## check
if not self.t_rec.getRef().equals( self.com.rec() )[1] or \
not self.t_lig.getRef().equals( self.com.lig() )[1]:
raise AnalyzeError('Atoms are not aligned.')
## native contact matrix
self.contacts = self.com.atomContacts()
self.hexContacts = None
'ldic':'pcr_lig/models.dic',
'rdic':'pcr_rec/models.dic'} )
if len( sys.argv ) < 2:
syntax()
fs = [ T.absfile( f ) for f in T.toList( options['i'] ) ]
result = ComplexList()
rec_dic = T.load( T.absfile( options['rdic'] ) )
lig_dic = T.load( T.absfile( options['ldic'] ) )
for f in fs:
T.flushPrint('Loading %s ...' % f )
cl = T.load( f )
cl = reduceComplexList( cl )
result += cl
T.flushPrint('done\n')
T.flushPrint('correct model numbers...')
correct_model_numbers( result, rec_dic, lig_dic )
T.flushPrint( '\ncasting all rec models...' )
pairwise_cast( result.models.recModels() )
###########################
# MAIN
###########################
if len(sys.argv) < 3:
_use()
options = t.cmdDict( defaultOptions() )
## ## current keys used for scoring
## scoreKeys = ['eProsa', 'ePairScore', 'foldX', 'cons_ent', 'cons_max']
## load docking solutions
t.flushPrint( "\nLoading complex list %s ... " % t.absfile( options['i'] ) )
complex_lst = t.load( options['i'] )
t.flushPrint( "done\n" )
## validate and expand list of keys to be calculated
force = []
if options.has_key('f'):
raw_force = t.toList( options['f'] )
## check that the key is valid
validKeys = ['fnac_4.5', 'fnac_10', 'fnrc_4.5',
'fnarc_9', 'fnarc_10', 'c_ratom_9', 'c_ratom_10',
'eProsa', 'ePairScore', 'foldX',
'cons_ent', 'cons_max', 'cons_abs',
'rms_if', 'rms_if_bb']
def randomSurfaces( base_folder, label, mask ):
"""
calculate surfaces for all peptides and return the
average and SD
"""
## container for results and standard deviations
MS, AS = {}, {}
MS_sd, AS_sd = {}, {}
## loop over peptide directories
for k in MOU.aaAtoms.keys():
dir = base_folder + 'GLY-%s-GLY_pcr/pcr_00'%(k)
fLst = glob.glob( dir + '/*.pdb')
msLst = []
asLst = []
## loop over pdb files for each peptide
T.flushPrint( '\nNow collecting data in %s'%dir )
for f in fLst:
## load peptide and remove waters and hydrogens
m = PDBModel( f )
m = m.compress( m.maskProtein() * m.maskHeavy() )
T.flushPrint( '.')
## add surface data
try:
d = PDBDope( m )
d.addSurfaceRacer( probe=1.4 )
## remove tailing GLY
m = m.compress( m.res2atomMask(mask) )
## collect surface data for each peptide
msLst += [ m.profile('MS') ]
asLst += [ m.profile('AS') ]
except:
print 'Failed calculating exposure for GLY-%s-GLY'%(k)
print '\t and file %s'%f
## get result dictionary for peptide
T.flushPrint('\nCollecting data ...\n')
msDic = {}
asDic = {}
msDic_sd = {}
asDic_sd = {}
j = 0
#atoms = [ a['name'] for a in m.atoms ]
for n in m['name']:
msDic[n] = N.average(msLst)[j]
asDic[n] = N.average(asLst)[j]
msDic_sd[n] = MAU.SD( msLst )[j]
asDic_sd[n] = MAU.SD( asLst )[j]
j += 1
MS[ k ] = msDic
AS[ k ] = asDic
MS_sd[ k ] = msDic_sd
AS_sd[ k ] = asDic_sd
return MS, AS, MS_sd, AS_sd
def retrievePDBs( self, outFolder=None, pdbCodes=None ):
"""
Get PDB from local database if it exists, if not try to
download the coordinartes drom the RSCB.
Write PDBs for given fasta records. Add PDB infos to internal
dictionary of fasta records. NMR structures get resolution 3.5.
@param outFolder: folder to put PDB files into (default: L{F_ALL})
@type outFolder: str OR None
@param pdbCodes: list of PDB codes [all previously found templates]
@type pdbCodes: [str]
@return: list of PDB file names
@rtype: [str]
@raise BlastError: if can't write file
"""
outFolder = outFolder or self.outFolder + self.F_ALL
pdbCodes = pdbCodes or self.record_dic.keys()
result = []
i = 0
if not self.silent:
T.flushPrint("fetching %i PDBs (l=local, r=remotely)..." % \
len( pdbCodes ) )
for c in pdbCodes:
i += 1
fname = '%s/%s.pdb' % (outFolder, c)
try:
if os.path.exists( fname ):
h = open( fname, 'r' )
else:
h = self.getLocalPDBHandle( c )
if not self.silent:
T.flushPrint('l')
except:
h = self.getRemotePDBHandle( c )
if not self.silent:
T.flushPrint('r')
try:
lines, infos = self.parsePdbFromHandle( h, first_model_only=1 )
infos['file'] = fname
if c in self.record_dic:
self.record_dic[ c ].__dict__.update( infos )
## close if it is a handle
try: h.close()
except:
pass
if not os.path.exists( fname ):
f = open( fname, 'w', 1 )
f.writelines( lines )
f.close()
result += [ fname ]
except IOError, why:
raise BlastError( "Can't write file "+fname )
