def wrapModelToNative(self,
mrPdb,
nativePdb,
origin=[0.0,0.0,0.0],
csymmatchPdb=None,
workdir=None,
cleanup=True):
"""Take a pdb and wrap it onto the nativePdb using csymmatch.
If origin is not [0.0,0.0,0.0] we also move the structure onto the new origin before wrapping"""
if workdir is None:
workdir = os.getcwd()
assert os.path.isfile(mrPdb) and os.path.isfile(nativePdb),"Cannot find: {0} or {1}".format(mrPdb,nativePdb)
originMrPdb = None
if origin != [ 0.0, 0.0, 0.0 ]:
ostr="o{}_{}".format(origin, str(uuid.uuid1())).replace(" ","" )
originMrPdb = ample_util.filename_append(filename=mrPdb, astr=ostr, directory=workdir)
pdb_edit.translate(inpdb=mrPdb, outpdb=originMrPdb, ftranslate=origin)
mrPdb = originMrPdb
if csymmatchPdb is None:
csymmatchPdb = ample_util.filename_append(filename=mrPdb, astr="csymmatch_{}".format(str(uuid.uuid1())),
directory=workdir)
self.run(refPdb=nativePdb, inPdb=mrPdb, outPdb=csymmatchPdb, originHand=False, cleanup=cleanup)
if not os.path.isfile( csymmatchPdb ):
raise RuntimeError("Error generating csymmatchPdb")
if cleanup and originMrPdb:
os.unlink(originMrPdb)
return csymmatchPdb
def run(nativePdb, nativeMtz=None, nativeMap=None, mrPdbs=None, outDir=None):
logging.info("Using native PDB file: %s", os.path.abspath(nativePdb))
logging.info("Using native MTZ file: %s", os.path.abspath(nativeMtz))
logging.info("Wrapping the following MTZ files to the native: %s",
" ".join(mrPdbs))
phenix = False
if outDir is not None:
if not os.path.isdir(outDir):
msg = "Cannot find output directory: {}".format(outDir)
raise RuntimeError(msg)
outDir = os.path.abspath(outDir)
else:
outDir = os.getcwd()
if phenix:
if nativeMap is None:
nativeMap = generateMap(nativePdb, nativeMtz)
if not os.path.isfile(nativeMap):
msg = "Cannot find nativeMap: {}".format(nativeMap)
raise RuntimeError(msg)
else:
shelxeExe = ample_util.find_exe('shelxe')
removables = []
for mrPdb in mrPdbs:
if phenix:
logging.debug("Searching for origin shift using: %s %s", nativeMap,
mrPdb)
origin = phenixer.ccmtzOrigin(nativeMap, mrPdb)
# offset.pdb is the mrPdb moved onto the new origin
offsetPdb = "offset.pdb"
logging.debug("Found origin: %s\nOffset pdb is: %s", origin,
offsetPdb)
else:
mrinfo = shelxe.MRinfo(shelxeExe, nativePdb, nativeMtz)
mrinfo.analyse(mrPdb)
originShift = mrinfo.originShift
logging.debug("Found origin: {0}".format(originShift))
offsetPdb = ample_util.filename_append(mrPdb,
astr='offset',
directory=os.getcwd())
pdb_edit.translate(mrPdb, offsetPdb, originShift)
csymmPdb = ample_util.filename_append(filename=mrPdb,
astr="csymmatch",
directory=outDir)
logging.debug("Running csymmatch to wrap %s onto native %s", offsetPdb,
nativePdb)
csymmatch.Csymmatch().run(refPdb=nativePdb,
inPdb=offsetPdb,
outPdb=csymmPdb,
originHand=False)
removables += [offsetPdb]
logging.info("Matched PDB is: %s", csymmPdb)
map(os.remove, removables + ["shelxe-input.hkl", "shelxe-input.ent"])
def scoreOrigin(self,
origin=None,
mrPdbInfo=None,
nativePdbInfo=None,
resSeqMap=None,
workdir=os.getcwd()
):
self.workdir = workdir
if not resSeqMap.resSeqMatch():
# We need to create a copy of the placed pdb with numbering matching the native
mrPdbRes = ample_util.filename_append( filename=mrPdbInfo.pdb, astr="reseq", directory=self.workdir )
pdb_edit.match_resseq( targetPdb=mrPdbInfo.pdb, sourcePdb=None, outPdb=mrPdbRes, resMap=resSeqMap )
mrPdb = mrPdbRes
else:
mrPdb = mrPdbInfo.pdb
# Make a copy of mrPdb with chains renamed to lower case
ucChains = mrPdbInfo.models[0].chains
toChains = [ c.lower() for c in ucChains ]
mrAaPdb = ample_util.filename_append( filename=mrPdb, astr="ren", directory=self.workdir )
pdb_edit.rename_chains( inpdb=mrPdb, outpdb=mrAaPdb, fromChain=ucChains, toChain=toChains )
# The list of chains in the native that we will be checking contacts from
fromChains = nativePdbInfo.models[0].chains
mrOriginPdb = mrAaPdb
if origin != [ 0.0, 0.0, 0.0 ]:
# Move pdb to new origin
#ostr="origin{0}".format(i)
ostr="o{0}".format( origin ).replace(" ","" )
mrOriginPdb = ample_util.filename_append( filename=mrAaPdb, astr=ostr, directory=self.workdir )
pdb_edit.translate( inpdb=mrAaPdb, outpdb=mrOriginPdb, ftranslate=origin )
# Concatenate into one file
joinedPdb = ample_util.filename_append( filename=mrOriginPdb, astr="joined", directory=self.workdir )
pdb_edit.merge( pdb1=nativePdbInfo.pdb, pdb2=mrOriginPdb, pdbout=joinedPdb )
# Run ncont
data = RioData()
data.origin = origin
data.originPdb = mrOriginPdb
data.joinedPdb = joinedPdb
data.fromChains = fromChains
data.toChains = toChains
# First get AllAtom score
self.calcAllAtom( data )
# Then score RIO
self.calcRio( data )
#data.numGood = data.inregister + data.ooregister
# clean up
os.unlink(mrOriginPdb)
os.unlink(joinedPdb)
if os.path.isfile(mrAaPdb): os.unlink(mrAaPdb)
return data
def analyseSolution(amoptd, d, mrinfo):
logger.info("Benchmark: analysing result: {0}".format(d['ensemble_name']))
mrPdb = None
if d['MR_program'] == "PHASER":
mrPdb = d['PHASER_pdbout']
mrMTZ = d['PHASER_mtzout']
elif d['MR_program'] == "MOLREP":
mrPdb = d['MOLREP_pdbout']
elif d['MR_program'] == "unknown":
return
if mrPdb is None or not os.path.isfile(mrPdb):
#logger.critical("Cannot find mrPdb {0} for solution {1}".format(mrPdb,d))
return
# debug - copy into work directory as reforigin struggles with long pathnames
shutil.copy(
mrPdb,
os.path.join(fixpath(amoptd['benchmark_dir']),
os.path.basename(mrPdb)))
mrPdbInfo = pdb_edit.get_info(mrPdb)
d['num_placed_chains'] = mrPdbInfo.numChains()
d['num_placed_atoms'] = mrPdbInfo.numAtoms()
d['num_placed_CA'] = mrPdbInfo.numCalpha()
if amoptd['native_pdb']:
if not d['SHELXE_os']:
logger.critical(
"mrPdb {0} has no SHELXE_os origin shift. Calculating...".
format(mrPdb))
mrinfo.analyse(mrPdb)
mrOrigin = mrinfo.originShift
d['SHELXE_MPE'] = mrinfo.MPE
d['SHELXE_wMPE'] = mrinfo.wMPE
else:
mrOrigin = [c * -1 for c in d['SHELXE_os']]
# Move pdb onto new origin
originPdb = ample_util.filename_append(mrPdb,
astr='offset',
directory=fixpath(
amoptd['benchmark_dir']))
#print(mrPdb, originPdb, mrOrigin)
pdb_edit.translate(mrPdb, originPdb, mrOrigin)
# offset.pdb is the mrModel shifted onto the new origin use csymmatch to wrap onto native
csymmatch.Csymmatch().wrapModelToNative(
originPdb,
amoptd['native_pdb'],
csymmatchPdb=os.path.join(
fixpath(amoptd['benchmark_dir']),
"phaser_{0}_csymmatch.pdb".format(d['ensemble_name'])))
# can now delete origin pdb
os.unlink(originPdb)
# Calculate phase error for the MR PDB
try:
mrinfo.analyse(mrPdb)
d['MR_MPE'] = mrinfo.MPE
d['MR_wMPE'] = mrinfo.wMPE
except Exception as e:
logger.critical("Error analysing mrPdb: {0}\n{1}".format(mrPdb, e))
# We cannot calculate the Reforigin RMSDs or RIO scores for runs where we don't have a full initial model
# to compare to the native to allow us to determine which parts of the ensemble correspond to which parts of
# the native structure.
if not (amoptd['homologs'] or \
amoptd['ideal_helices'] or \
amoptd['import_ensembles'] or \
amoptd['single_model_mode']):
# Get reforigin info
rmsder = reforigin.ReforiginRmsd()
try:
rmsder.getRmsd(nativePdbInfo=amoptd['native_pdb_info'],
placedPdbInfo=mrPdbInfo,
refModelPdbInfo=amoptd['ref_model_pdb_info'],
cAlphaOnly=True,
workdir=fixpath(amoptd['benchmark_dir']))
d['reforigin_RMSD'] = rmsder.rmsd
except Exception as e:
logger.critical("Error calculating RMSD: {0}".format(e))
d['reforigin_RMSD'] = 999
# Score the origin with all-atom and rio
rioData = rio.Rio().scoreOrigin(
mrOrigin,
mrPdbInfo=mrPdbInfo,
nativePdbInfo=amoptd['native_pdb_info'],
resSeqMap=amoptd['res_seq_map'],
workdir=fixpath(amoptd['benchmark_dir']))
# Set attributes
d['AA_num_contacts'] = rioData.aaNumContacts
d['RIO_num_contacts'] = rioData.rioNumContacts
d['RIO_in_register'] = rioData.rioInRegister
d['RIO_oo_register'] = rioData.rioOoRegister
d['RIO_backwards'] = rioData.rioBackwards
d['RIO'] = rioData.rioInRegister + rioData.rioOoRegister
d['RIO_no_cat'] = rioData.rioNumContacts - (rioData.rioInRegister +
rioData.rioOoRegister)
d['RIO_norm'] = float(d['RIO']) / float(
d['native_pdb_num_residues'])
else:
d['AA_num_contacts'] = None
d['RIO_num_contacts'] = None
d['RIO_in_register'] = None
d['RIO_oo_register'] = None
d['RIO_backwards'] = None
d['RIO'] = None
d['RIO_no_cat'] = None
d['RIO_norm'] = None
# # Now get the helix
# helixSequence = contacts.Rio().helixFromContacts( contacts=rioData.contacts,
# dsspLog=dsspLog )
# if helixSequence is not None:
# ampleResult.rioHelixSequence = helixSequence
# ampleResult.rioLenHelix = len( helixSequence )
# hfile = os.path.join( workdir, "{0}.helix".format( ampleResult.ensembleName ) )
# with open( hfile, 'w' ) as f:
# f.write( helixSequence+"\n" )
#
# This purely for checking and so we have pdbs to view
#
# Wrap shelxe trace onto native using Csymmatch
if not d['SHELXE_pdbout'] is None and os.path.isfile(
fixpath(d['SHELXE_pdbout'])):
csymmatch.Csymmatch().wrapModelToNative(
fixpath(d['SHELXE_pdbout']),
amoptd['native_pdb'],
origin=mrOrigin,
workdir=fixpath(amoptd['benchmark_dir']))
if not ('SHELXE_wMPE' in d and d['SHELXE_wMPE']):
try:
mrinfo.analyse(d['SHELXE_pdbout'])
d['SHELXE_MPE'] = mrinfo.MPE
d['SHELXE_wMPE'] = mrinfo.wMPE
except Exception as e:
logger.critical(
"Error analysing SHELXE_pdbout: {0}\n{1}".format(
d['SHELXE_pdbout'], e))
# Wrap parse_buccaneer model onto native
if d['SXRBUCC_pdbout'] and os.path.isfile(fixpath(
d['SXRBUCC_pdbout'])):
# Need to rename Pdb as is just called buccSX_output.pdb
csymmatchPdb = os.path.join(
fixpath(amoptd['benchmark_dir']),
"buccaneer_{0}_csymmatch.pdb".format(d['ensemble_name']))
csymmatch.Csymmatch().wrapModelToNative(
fixpath(d['SXRBUCC_pdbout']),
amoptd['native_pdb'],
origin=mrOrigin,
csymmatchPdb=csymmatchPdb,
workdir=fixpath(amoptd['benchmark_dir']))
# Calculate phase error
try:
mrinfo.analyse(d['SXRBUCC_pdbout'])
d['SXRBUCC_MPE'] = mrinfo.MPE
d['SXRBUCC_wMPE'] = mrinfo.wMPE
except Exception as e:
logger.critical(
"Error analysing SXRBUCC_pdbout: {0}\n{1}".format(
d['SXRBUCC_pdbout'], e))
# Wrap parse_buccaneer model onto native
if d['SXRARP_pdbout'] and os.path.isfile(fixpath(d['SXRARP_pdbout'])):
# Need to rename Pdb as is just called buccSX_output.pdb
csymmatchPdb = os.path.join(
fixpath(amoptd['benchmark_dir']),
"arpwarp_{0}_csymmatch.pdb".format(d['ensemble_name']))
csymmatch.Csymmatch().wrapModelToNative(
fixpath(d['SXRARP_pdbout']),
amoptd['native_pdb'],
origin=mrOrigin,
csymmatchPdb=csymmatchPdb,
workdir=fixpath(amoptd['benchmark_dir']))
# Calculate phase error
try:
mrinfo.analyse(d['SXRARP_pdbout'])
d['SXRARP_MPE'] = mrinfo.MPE
d['SXRARP_wMPE'] = mrinfo.wMPE
except Exception as e:
logger.critical(
"Error analysing SXRARP_pdbout: {0}\n{1}".format(
d['SXRARP_pdbout'], e))
return
def scoreOrigin(self,
origin=None,
mrPdbInfo=None,
nativePdbInfo=None,
resSeqMap=None,
workdir=os.getcwd()):
self.workdir = workdir
if not resSeqMap.resSeqMatch():
# We need to create a copy of the placed pdb with numbering matching the native
mrPdbRes = ample_util.filename_append(filename=mrPdbInfo.pdb,
astr="reseq",
directory=self.workdir)
pdb_edit.match_resseq(targetPdb=mrPdbInfo.pdb,
sourcePdb=None,
outPdb=mrPdbRes,
resMap=resSeqMap)
mrPdb = mrPdbRes
else:
mrPdb = mrPdbInfo.pdb
# Make a copy of mrPdb with chains renamed to lower case
ucChains = mrPdbInfo.models[0].chains
toChains = [c.lower() for c in ucChains]
mrAaPdb = ample_util.filename_append(filename=mrPdb,
astr="ren",
directory=self.workdir)
pdb_edit.rename_chains(inpdb=mrPdb,
outpdb=mrAaPdb,
fromChain=ucChains,
toChain=toChains)
# The list of chains in the native that we will be checking contacts from
fromChains = nativePdbInfo.models[0].chains
mrOriginPdb = mrAaPdb
if origin != [0.0, 0.0, 0.0]:
# Move pdb to new origin
#ostr="origin{0}".format(i)
ostr = "o{0}".format(origin).replace(" ", "")
mrOriginPdb = ample_util.filename_append(filename=mrAaPdb,
astr=ostr,
directory=self.workdir)
pdb_edit.translate(inpdb=mrAaPdb,
outpdb=mrOriginPdb,
ftranslate=origin)
# Concatenate into one file
joinedPdb = ample_util.filename_append(filename=mrOriginPdb,
astr="joined",
directory=self.workdir)
pdb_edit.merge(pdb1=nativePdbInfo.pdb,
pdb2=mrOriginPdb,
pdbout=joinedPdb)
# Run ncont
data = RioData()
data.origin = origin
data.originPdb = mrOriginPdb
data.joinedPdb = joinedPdb
data.fromChains = fromChains
data.toChains = toChains
# First get AllAtom score
self.calcAllAtom(data)
# Then score RIO
self.calcRio(data)
#data.numGood = data.inregister + data.ooregister
# clean up
os.unlink(mrOriginPdb)
os.unlink(joinedPdb)
if os.path.isfile(mrAaPdb): os.unlink(mrAaPdb)
return data
def findOrigin(self,
nativePdbInfo=None,
mrPdbInfo=None,
resSeqMap=None,
origins=None,
allAtom=False,
workdir=os.getcwd()):
"""Find the origin using the maximum number of contacts as metric"""
self.workdir = workdir
if not resSeqMap.resSeqMatch():
# We need to create a copy of the placed pdb with numbering matching the native
mrPdbRes = ample_util.filename_append(filename=mrPdbInfo.pdb,
astr="reseq",
directory=self.workdir)
pdb_edit.match_resseq(targetPdb=mrPdbInfo.pdb,
sourcePdb=None,
outPdb=mrPdbRes,
resMap=resSeqMap)
mrPdb = mrPdbRes
else:
mrPdb = mrPdbInfo.pdb
# Make a copy of mrPdb with chains renamed to lower case
ucChains = mrPdbInfo.models[0].chains
toChains = [c.lower() for c in ucChains]
placedAaPdb = ample_util.filename_append(filename=mrPdb,
astr="ren",
directory=self.workdir)
pdb_edit.rename_chains(inpdb=mrPdb,
outpdb=placedAaPdb,
fromChain=ucChains,
toChain=toChains)
# The list of chains in the native that we will be checking contacts from
fromChains = nativePdbInfo.models[0].chains
# Loop over origins, move the placed pdb to the new origin and then run ncont
# Object to hold data on best origin
self.data = None
for origin in origins:
placedOriginPdb = placedAaPdb
if origin != [0.0, 0.0, 0.0]:
# Move pdb to new origin
#ostr="origin{0}".format(i)
ostr = "o{0}".format(origin).replace(" ", "")
placedOriginPdb = ample_util.filename_append(
filename=placedAaPdb, astr=ostr, directory=self.workdir)
pdb_edit.translate(inpdb=placedAaPdb,
outpdb=placedOriginPdb,
ftranslate=origin)
# Concatenate into one file
joinedPdb = ample_util.filename_append(filename=placedOriginPdb,
astr="joined",
directory=self.workdir)
pdb_edit.merge(pdb1=nativePdbInfo.pdb,
pdb2=placedOriginPdb,
pdbout=joinedPdb)
# Set up object to hold data
data = RioData()
data.origin = origin
data.originPdb = placedOriginPdb
data.joinedPdb = joinedPdb
data.fromChains = fromChains
data.toChains = toChains
data.numGood = 0 # For holding the metric
# Run ncont
if allAtom:
self.calcAllAtom(data)
data.numGood = data.aaNumContacts
else:
self.calcRio(data)
data.numGood = data.rioInRegister + data.rioOoRegister
# Save the first origin and only update if we get a better score
if not self.data or data.numGood > self.data.numGood:
self.data = data
# End loop over origins
# Now need to calculate data for whichever one we didn't calculate
if allAtom:
self.calcRio(self.data)
else:
self.calcAllAtom(self.data)
if self.data.numGood > 0:
# If we got a match run csymmatch so we can see the result
csym = csymmatch.Csymmatch()
csymmatchPdb = ample_util.filename_append(
filename=self.data.originPdb,
astr="csymmatch_best",
directory=self.workdir)
csym.run(refPdb=nativePdbInfo.pdb,
inPdb=self.data.originPdb,
outPdb=csymmatchPdb,
originHand=False)
return self.data
def analyseSolution(amoptd, d, mrinfo):
logger.info("Benchmark: analysing result: {0}".format(d['ensemble_name']))
mrPdb=None
if d['MR_program']=="PHASER":
mrPdb = d['PHASER_pdbout']
mrMTZ = d['PHASER_mtzout']
elif d['MR_program']=="MOLREP":
mrPdb = d['MOLREP_pdbout']
elif d['MR_program']=="unknown":
return
if mrPdb is None or not os.path.isfile(mrPdb):
#logger.critical("Cannot find mrPdb {0} for solution {1}".format(mrPdb,d))
return
# debug - copy into work directory as reforigin struggles with long pathnames
shutil.copy(mrPdb, os.path.join(fixpath(amoptd['benchmark_dir']), os.path.basename(mrPdb)))
mrPdbInfo = pdb_edit.get_info( mrPdb )
d['num_placed_chains'] = mrPdbInfo.numChains()
d['num_placed_atoms'] = mrPdbInfo.numAtoms()
d['num_placed_CA'] = mrPdbInfo.numCalpha()
if amoptd['native_pdb']:
if not d['SHELXE_os']:
logger.critical("mrPdb {0} has no SHELXE_os origin shift. Calculating...".format(mrPdb))
mrinfo.analyse(mrPdb)
mrOrigin = mrinfo.originShift
d['SHELXE_MPE'] = mrinfo.MPE
d['SHELXE_wMPE'] = mrinfo.wMPE
else:
mrOrigin=[c*-1 for c in d['SHELXE_os']]
# Move pdb onto new origin
originPdb = ample_util.filename_append(mrPdb, astr='offset',directory=fixpath(amoptd['benchmark_dir']))
#print(mrPdb, originPdb, mrOrigin)
pdb_edit.translate(mrPdb, originPdb, mrOrigin)
# offset.pdb is the mrModel shifted onto the new origin use csymmatch to wrap onto native
csymmatch.Csymmatch().wrapModelToNative(originPdb,
amoptd['native_pdb'],
csymmatchPdb=os.path.join(fixpath(amoptd['benchmark_dir']),
"phaser_{0}_csymmatch.pdb".format(d['ensemble_name'])))
# can now delete origin pdb
os.unlink(originPdb)
# Calculate phase error for the MR PDB
try:
mrinfo.analyse(mrPdb)
d['MR_MPE'] = mrinfo.MPE
d['MR_wMPE'] = mrinfo.wMPE
except Exception as e:
logger.critical("Error analysing mrPdb: {0}\n{1}".format(mrPdb,e))
# We cannot calculate the Reforigin RMSDs or RIO scores for runs where we don't have a full initial model
# to compare to the native to allow us to determine which parts of the ensemble correspond to which parts of
# the native structure - or if we were unable to calculate a res_seq_map
if not (amoptd['homologs'] or \
amoptd['ideal_helices'] or \
amoptd['import_ensembles'] or \
amoptd['single_model_mode'] or \
amoptd['res_seq_map']):
# Get reforigin info
rmsder = reforigin.ReforiginRmsd()
try:
rmsder.getRmsd(nativePdbInfo=amoptd['native_pdb_info'],
placedPdbInfo=mrPdbInfo,
refModelPdbInfo=amoptd['ref_model_pdb_info'],
cAlphaOnly=True,
workdir=fixpath(amoptd['benchmark_dir']))
d['reforigin_RMSD'] = rmsder.rmsd
except Exception as e:
logger.critical("Error calculating RMSD: {0}".format(e))
d['reforigin_RMSD'] = 999
# Score the origin with all-atom and rio
rioData = rio.Rio().scoreOrigin(mrOrigin,
mrPdbInfo=mrPdbInfo,
nativePdbInfo=amoptd['native_pdb_info'],
resSeqMap=amoptd['res_seq_map'],
workdir=fixpath(amoptd['benchmark_dir'])
)
# Set attributes
d['AA_num_contacts'] = rioData.aaNumContacts
d['RIO_num_contacts'] = rioData.rioNumContacts
d['RIO_in_register'] = rioData.rioInRegister
d['RIO_oo_register'] = rioData.rioOoRegister
d['RIO_backwards'] = rioData.rioBackwards
d['RIO'] = rioData.rioInRegister + rioData.rioOoRegister
d['RIO_no_cat'] = rioData.rioNumContacts - ( rioData.rioInRegister + rioData.rioOoRegister )
d['RIO_norm'] = float(d['RIO']) / float(d['native_pdb_num_residues'])
else:
d['AA_num_contacts'] = None
d['RIO_num_contacts'] = None
d['RIO_in_register'] = None
d['RIO_oo_register'] = None
d['RIO_backwards'] = None
d['RIO'] = None
d['RIO_no_cat'] = None
d['RIO_norm'] = None
# # Now get the helix
# helixSequence = contacts.Rio().helixFromContacts( contacts=rioData.contacts,
# dsspLog=dsspLog )
# if helixSequence is not None:
# ampleResult.rioHelixSequence = helixSequence
# ampleResult.rioLenHelix = len( helixSequence )
# hfile = os.path.join( workdir, "{0}.helix".format( ampleResult.ensembleName ) )
# with open( hfile, 'w' ) as f:
# f.write( helixSequence+"\n" )
#
# This purely for checking and so we have pdbs to view
#
# Wrap shelxe trace onto native using Csymmatch
if not d['SHELXE_pdbout'] is None and os.path.isfile(fixpath(d['SHELXE_pdbout'])):
csymmatch.Csymmatch().wrapModelToNative( fixpath(d['SHELXE_pdbout']),
amoptd['native_pdb'],
origin=mrOrigin,
workdir=fixpath(amoptd['benchmark_dir']))
if not('SHELXE_wMPE' in d and d['SHELXE_wMPE']):
try:
mrinfo.analyse(d['SHELXE_pdbout'])
d['SHELXE_MPE'] = mrinfo.MPE
d['SHELXE_wMPE'] = mrinfo.wMPE
except Exception as e:
logger.critical("Error analysing SHELXE_pdbout: {0}\n{1}".format(d['SHELXE_pdbout'],e))
# Wrap parse_buccaneer model onto native
if d['SXRBUCC_pdbout'] and os.path.isfile(fixpath(d['SXRBUCC_pdbout'])):
# Need to rename Pdb as is just called buccSX_output.pdb
csymmatchPdb = os.path.join(fixpath(amoptd['benchmark_dir']), "buccaneer_{0}_csymmatch.pdb".format(d['ensemble_name']))
csymmatch.Csymmatch().wrapModelToNative( fixpath(d['SXRBUCC_pdbout']),
amoptd['native_pdb'],
origin=mrOrigin,
csymmatchPdb=csymmatchPdb,
workdir=fixpath(amoptd['benchmark_dir']))
# Calculate phase error
try:
mrinfo.analyse(d['SXRBUCC_pdbout'])
d['SXRBUCC_MPE'] = mrinfo.MPE
d['SXRBUCC_wMPE'] = mrinfo.wMPE
except Exception as e:
logger.critical("Error analysing SXRBUCC_pdbout: {0}\n{1}".format(d['SXRBUCC_pdbout'],e))
# Wrap parse_buccaneer model onto native
if d['SXRARP_pdbout'] and os.path.isfile(fixpath(d['SXRARP_pdbout'])):
# Need to rename Pdb as is just called buccSX_output.pdb
csymmatchPdb = os.path.join(fixpath(amoptd['benchmark_dir']), "arpwarp_{0}_csymmatch.pdb".format(d['ensemble_name']))
csymmatch.Csymmatch().wrapModelToNative( fixpath(d['SXRARP_pdbout']),
amoptd['native_pdb'],
origin=mrOrigin,
csymmatchPdb=csymmatchPdb,
workdir=fixpath(amoptd['benchmark_dir']))
# Calculate phase error
try:
mrinfo.analyse(d['SXRARP_pdbout'])
d['SXRARP_MPE'] = mrinfo.MPE
d['SXRARP_wMPE'] = mrinfo.wMPE
except Exception as e:
logger.critical("Error analysing SXRARP_pdbout: {0}\n{1}".format(d['SXRARP_pdbout'],e))
return
def findOrigin(self,
nativePdbInfo=None,
mrPdbInfo=None,
resSeqMap=None,
origins=None,
allAtom=False,
workdir=os.getcwd() ):
"""Find the origin using the maximum number of contacts as metric"""
self.workdir = workdir
if not resSeqMap.resSeqMatch():
# We need to create a copy of the placed pdb with numbering matching the native
mrPdbRes = ample_util.filename_append( filename=mrPdbInfo.pdb, astr="reseq", directory=self.workdir )
pdb_edit.match_resseq( targetPdb=mrPdbInfo.pdb, sourcePdb=None, outPdb=mrPdbRes, resMap=resSeqMap )
mrPdb = mrPdbRes
else:
mrPdb = mrPdbInfo.pdb
# Make a copy of mrPdb with chains renamed to lower case
ucChains = mrPdbInfo.models[0].chains
toChains = [ c.lower() for c in ucChains ]
placedAaPdb = ample_util.filename_append( filename=mrPdb, astr="ren", directory=self.workdir )
pdb_edit.rename_chains( inpdb=mrPdb, outpdb=placedAaPdb, fromChain=ucChains, toChain=toChains )
# The list of chains in the native that we will be checking contacts from
fromChains = nativePdbInfo.models[0].chains
# Loop over origins, move the placed pdb to the new origin and then run ncont
# Object to hold data on best origin
self.data = None
for origin in origins:
placedOriginPdb = placedAaPdb
if origin != [ 0.0, 0.0, 0.0 ]:
# Move pdb to new origin
#ostr="origin{0}".format(i)
ostr="o{0}".format( origin ).replace(" ","" )
placedOriginPdb = ample_util.filename_append( filename=placedAaPdb, astr=ostr, directory=self.workdir )
pdb_edit.translate( inpdb=placedAaPdb, outpdb=placedOriginPdb, ftranslate=origin )
# Concatenate into one file
joinedPdb = ample_util.filename_append( filename=placedOriginPdb, astr="joined", directory=self.workdir )
pdb_edit.merge( pdb1=nativePdbInfo.pdb, pdb2=placedOriginPdb, pdbout=joinedPdb )
# Set up object to hold data
data = RioData()
data.origin = origin
data.originPdb = placedOriginPdb
data.joinedPdb = joinedPdb
data.fromChains = fromChains
data.toChains = toChains
data.numGood = 0 # For holding the metric
# Run ncont
if allAtom:
self.calcAllAtom( data )
data.numGood = data.aaNumContacts
else:
self.calcRio( data )
data.numGood = data.rioInRegister + data.rioOoRegister
# Save the first origin and only update if we get a better score
if not self.data or data.numGood > self.data.numGood:
self.data = data
# End loop over origins
# Now need to calculate data for whichever one we didn't calculate
if allAtom:
self.calcRio( self.data )
else:
self.calcAllAtom( self.data )
if self.data.numGood > 0:
# If we got a match run csymmatch so we can see the result
csym = csymmatch.Csymmatch()
csymmatchPdb = ample_util.filename_append( filename=self.data.originPdb,
astr="csymmatch_best",
directory=self.workdir )
csym.run( refPdb=nativePdbInfo.pdb,
inPdb=self.data.originPdb,
outPdb=csymmatchPdb,
originHand=False )
return self.data
