def rundmp(spath, msapath, outdir):
"""Run DeepMetaPSICOV to produce contact prediction lists.
:param seqpath: Input sequence filepath.
:type seqpath: str
:param msapath: Input MSA filepath.
:type msapath: str
"""
dmp_exec = '"' + ppaths.check_path(DMP_PATH, 'file') + '"'
try:
os.system(dmp_exec + ' -i ' + spath + ' -a ' + msapath)
except Exception:
logging.critical(
' An error occurred while executing DeepMetaPSICOV.')
raise OSError
def _check_uniprot(inuniprot):
"""Return Uniprot segment threshold value and Uniprot database path.
:param inuniprot: Initial argument for Uniprot threshold
:type inuniprot: str
:raises ValueError: Argument is not valid.
:return: Threshold value and database path.
:rtype: float, str
"""
try:
float(inuniprot)
except Exception:
logging.critical('Uniprot threshold given not valid.')
raise ValueError
try:
dbpath = check_path(pconf.UNICLUST_FASTA_PATH, 'file')
except Exception:
logging.warning(
'Uniprot database file does not exist. Switching to server-only.')
dbpath = 'server-only'
return float(inuniprot), dbpath
def main():
parser = create_argument_parser()
args = parser.parse_args()
global logger
logger = pcl.pisacov_logger(level="info")
welcomemsg, starttime = pcl.welcome(command=__script__)
logger.info(welcomemsg)
kwlist = pco._default_keys()
configin = {}
if args.update_sifts_database is not None:
outpath = ppaths.check_path(args.update_sifts_database[0], 'either')
if os.path.isdir(outpath) is True:
outpath = os.path.join(outpath,
('pdb_chain_uniprot' + os.extsep + 'csv'))
pol.getsifts(outpath)
configin['SIFTS_PATH'] = outpath
else:
pass
if args.view_configuration is True:
print('** ' + __prog__ + " v." + __version__ +
' configuration file **' + os.linesep)
with open(pconf.__file__) as f:
for line in f:
line = _lineformat(line)
print(line, end='')
print('** End of configuration file **' + os.linesep)
return
else:
pass
if args.get_confpath is True:
print(pconf.__file__)
return
else:
pass
if args.conf_file is False:
newconf = pco._parse_conf()
else:
newconf = {}
if args.reset_hhblits_arguments is True:
newconf['HHBLITS_PARAMETERS'] = pco._default_values(
'HHBLITS_PARAMETERS')
else:
pass
if args.sifts_path is not None:
configin['SIFTS_PATH'] = args.sifts_path[0]
if args.pisa_path is not None:
configin['PISA_PATH'] = args.pisa_path[0]
if args.hhblits_path is not None:
configin['HHBLITS_PATH'] = args.hhblits_path[0]
if args.dmp_path is not None:
configin['DMP_PATH'] = args.dmp_path[0]
if args.uniclust_path is not None:
configin['UNICLUST_FASTA_PATH'] = args.uniclust_path[0]
if args.hhblits_arguments is not None:
configin['HHBLITS_PARAMETERS'] = args.hhblits_arguments[0]
if args.neighbours is not None:
configin['NEIGHBOURS_MINDISTANCE'] = args.neighbours[0]
configin['REMOVE_INTRA_CONTACTS'] = False
if isinstance(args.hhblits_location, list) is True:
configin['HHBLITS_DATABASE_NAME'] = args.hhblits_location[0]
configin['HHBLITS_DATABASE_DIR'] = args.hhblits_location[1]
compmsg = {
'SIFTS_PATH':
"Please, enter the path to the SIFTS csv file:\n",
'PISA_PATH':
"Please, enter the path to the PISA executable:\n",
'HHBLITS_PATH':
"Please, enter the path to the HHBLITS executable:\n",
'HHBLITS_DATABASE_NAME':
("Please, enter the name of the HHBLITS database name\n" +
"as requested by DeepMetaPSICOV (e.g. uniclust30_2018_08):\n"),
'HHBLITS_DATABASE_DIR':
"Please, enter the directory of the HHBLITS database:\n",
'DMP_PATH':
"Please, enter the path to the DeepMetaPSICOV executable:\n",
'HHBLITS_PARAMETERS':
("Please, enter the 5 HHBLITS parameters.\n" +
"#iterations, E-value cutoff, Non-redundant seqs to keep, " +
" MinimumCoverageWithMasterSeq(%), and MaxPairwiseSequenceIdentity.\n"
+ "Leave empty for DMP default (3, 0.001, 'inf', 50, 99).\n"),
'UNICLUST_FASTA_PATH':
("Please, enter the path to the UniClust fasta file.\n"
"An empty input will deactivate this option.\n"),
'NEIGHBOURS_MINDISTANCE':
("Press ENTER for intramolecular contacts " +
"to be removed from intermolecular contact lists.\n" +
"Otherwise, enter the minimum distance to be cosidered " +
"between neighbours. This will override the removal of" +
"intramolecular contacts that will be now ignored.\n")
}
for keystr in kwlist[:-1]:
if keystr in configin or args.conf_file is True:
if args.conf_file is True:
while True:
newval = input(compmsg[keystr])
try:
newconf[keystr] = pco._check_input(newval, keystr)
if keystr == 'NEIGHBOURS_MINDISTANCE':
if newval is None or newval == "":
newconf['REMOVE_INTRA_CONTACTS'] = True
else:
newconf['REMOVE_INTRA_CONTACTS'] = False
except Exception:
print("Not a valid input. Please, try again:")
else:
break
else:
newconf[keystr] = pco._check_input(configin[keystr], keystr)
if keystr == 'NEIGHBOURS_MINDISTANCE':
newconf['REMOVE_INTRA_CONTACTS'] = pco._check_input(
configin['REMOVE_INTRA_CONTACTS'],
'REMOVE_INTRA_CONTACTS')
else:
pass
if 'REMOVE_INTRA_CONTACTS' in configin:
newconf['REMOVE_INTRA_CONTACTS'] = pco._check_input(
configin['REMOVE_INTRA_CONTACTS'], 'REMOVE_INTRA_CONTACTS')
if newconf['REMOVE_INTRA_CONTACTS'] is True:
newconf['NEIGHBOURS_MINDISTANCE'] == 2
else:
pass
else:
pass
_outconffile(newconf)
endmsg = pcl.ok(starttime, command=__script__)
logger.info(endmsg)
return
def main():
parser = create_argument_parser()
args = parser.parse_args()
global logger
logger = pcl.pisacov_logger(level="info")
welcomemsg, starttime = pcl.welcome(command=__script__)
logger.info(welcomemsg)
# PARSE CONFIGURATION FILE:
invals = pco._initialise_inputs()
invals['INSEQ'] = None
invals['INSTR'] = None
invals['ALTDB'] = None
invals['OUTROOT'] = None
invals['OUTCSVPATH'] = None
invals['UPTHRESHOLD'] = None
# READ INPUT ARGUMENTS
invals['INSEQ'] = ppaths.check_path(args.seqpath[0], 'file')
invals['INSTR'] = ppaths.check_path(args.crystalpath[0], 'file')
if args.hhblits_arguments is not None:
invals['HHBLITS_PARAMETERS'] = pco._check_hhparams(
args.hhblits_arguments)
else:
pass
if args.uniprot_threshold is not None:
try:
invals['UPTHRESHOLD'] = float(args.uniprot_threshold[0])
except ValueError:
logger.critical('Uniprot threshold given not valid.')
if invals['UNICLUST_FASTA_PATH'] is None:
invals['UNICLUST_FASTA_PATH'] = pco._uniurl
else:
pass
if args.skip_conpred is True:
skipexec = True
if (args.hhblits_arguments is not None
or args.uniprot_threshold is not None):
logger.info(
'HHblits, UniProt threshold parameters given bypassed by --skip_conpred'
)
else:
skipexec = False
cropping = args.remove_insertions
scoring = [cropping, not cropping]
if args.outdir is None:
invals['OUTROOT'] = ppaths.check_path(os.path.dirname(invals['INSEQ']))
else:
invals['OUTROOT'] = ppaths.check_path(os.path.join(args.outdir[0], ''))
ppaths.mdir(invals['OUTROOT'])
invals['OUTCSVPATH'] = []
if args.collection_file is None:
invals['OUTCSVPATH'].append(
ppaths.check_path(
os.path.join(invals['OUTROOT'],
("evcovsignal" + os.extsep + "cropped" +
os.extsep + "pisacov" + os.extsep + "csv"))))
invals['OUTCSVPATH'].append(
ppaths.check_path(
os.path.join(invals['OUTROOT'],
("evcovsignal" + os.extsep + "full" + os.extsep +
"pisacov" + os.extsep + "csv"))))
else:
if cropping is True:
invals['OUTCSVPATH'].append(
ppaths.check_path(args.collection_file[0]))
invals['OUTCSVPATH'].append(
ppaths.check_path(
os.path.splitext(args.collection_file[0])[0] + os.extsep +
'full' + os.extsep +
os.path.splitext(args.collection_file[0])[1]))
else:
invals['OUTCSVPATH'].append(None)
invals['OUTCSVPATH'].append(
ppaths.check_path(args.collection_file[0]))
# Define formats used
sources = pco._sources()
# Parse sequence and structure files
logger.info('Parsing sequence file...')
seqs = cps.parseseqfile(invals['INSEQ'])
logger.info('Parsing structure file...')
strs, filestrs = cps.parsestrfile(invals['INSTR'])
if len(seqs) == 1 or len(strs) == 1:
if len(seqs) == 1:
for key in seqs:
pdbid = key
elif len(seqs) > 1 and len(strs) == 1:
for key in strs:
for key2 in seqs:
if key.upper() == key2.upper():
pdbid = key.upper()
else:
if key2.upper() in key.upper():
pdbid = key2.upper()
else:
raise Exception(
'More than one pdbid in sequence and/or structure set.')
seq = seqs[pdbid]
#structure = strs[pdbid]
# CROPPING AND RENUMBERING
outpdbdir = os.path.join(invals['OUTROOT'], pdbid, "")
instrc = os.path.join(invals['OUTROOT'], pdbid,
os.path.basename(invals['INSTR']))
fseq = {}
fmsa = {}
if skipexec is False:
if cropping is True:
logger.info('Cropping and renumbering sequences, ' +
'structures according to SIFTS database.')
logger.info(pcl.running('CROPS-cropstr'))
itime = datetime.datetime.now()
psc.runcrops(invals['INSEQ'], invals['INSTR'],
invals['SIFTS_PATH'], invals['UPTHRESHOLD'],
invals['UNICLUST_FASTA_PATH'], invals['OUTROOT'])
logger.info(pcl.running('CROPS-cropstr', done=itime))
else:
logger.info('Renumbering structure ' +
'according to position in sequence.')
logger.info(pcl.running('CROPS-renumber'))
itime = datetime.datetime.now()
psc.renumcrops(invals['INSEQ'], invals['INSTR'], invals['OUTROOT'])
logger.info(pcl.running('CROPS-renumber', done=itime))
ppaths.mdir(outpdbdir)
if cropping is False:
psc.splitseqs(invals['INSEQ'], outpdbdir)
copyfile(invals['INSTR'], instrc)
for i, iseq in seq.imer.items():
fiseq = pdbid + '_' + i + '.fasta'
fseq[i] = os.path.join(invals['OUTROOT'], pdbid, fiseq)
fiseq = pdbid + '_' + i + '.msa.aln'
fmsa[i] = os.path.join(invals['OUTROOT'], pdbid, 'hhblits', fiseq)
if skipexec is False:
iseq.dump(fseq[i])
# Parse cropped sequences and maps
if cropping is True:
amap = {}
fcropseq = {}
fcropmsa = {}
for i, iseq in seq.imer.items():
fprefix = pdbid + '_' + i + '.crops.to_uniprot'
fmap = os.path.join(invals['OUTROOT'], pdbid,
fprefix + os.extsep + 'cropmap')
amap.update(cps.parsemapfile(fmap)[pdbid])
fcropseq[i] = os.path.join(invals['OUTROOT'], pdbid,
fprefix + os.extsep + 'fasta')
fcropmsa[i] = os.path.join(
invals['OUTROOT'], pdbid, 'hhblits',
(fprefix + os.extsep + 'msa' + os.extsep + 'aln'))
seq.set_cropmaps(amap, cropmain=True)
# EXECUTION OF EXTERNAL PROGRAMS
hhdir = os.path.join(invals['OUTROOT'], pdbid, 'hhblits', '')
dmpdir = os.path.join(invals['OUTROOT'], pdbid, 'dmp', '')
pisadir = os.path.join(invals['OUTROOT'], pdbid, 'pisa', '')
fstr = os.path.join(
invals['OUTROOT'],
(pdbid + os.extsep + 'crops' + os.extsep + 'seq' + os.extsep + 'pdb'))
if cropping:
fcropstr = os.path.join(
invals['OUTROOT'], pdbid,
(pdbid + os.extsep + 'crops' + os.extsep + 'oldids' + os.extsep +
'to_uniprot' + os.path.splitext(invals['INSTR'])[1]))
if skipexec is False:
# MSA GENERATOR
ppaths.mdir(hhdir)
if invals['HHBLITS_PARAMETERS'] == ['3', '0.001', 'inf', '50', '99']:
logger.info(
'Generating Multiple Sequence Alignment using DeepMetaPSICOV default parameters... [AS RECOMMENDED]'
)
elif invals['HHBLITS_PARAMETERS'] == ['2', '0.001', '1000', '0', '90']:
logger.info(
'Generating Multiple Sequence Alignment using HHBlits default parameters...'
)
else:
logger.info(
'Generating Multiple Sequence Alignment using user-custom parameters...'
)
for i, iseq in seq.imer.items():
sfile = fcropseq[i] if cropping is True else fseq[i]
afile = fcropmsa[i] if cropping is True else fmsa[i]
logger.info(pcl.running('HHBlits'))
itime = datetime.datetime.now()
themsa = psm.runhhblits(sfile, invals['HHBLITS_PARAMETERS'], hhdir)
logger.info(pcl.running('HHBlits', done=itime))
if cropping is True:
iseq.cropmsa = themsa
if iseq.ncrops() == 0:
iseq.msa = iseq.cropmsa
logger.info(' Cropped sequence ' + iseq.oligomer_id +
'_' + iseq.name +
' is identical to original sequence.')
continue
else:
pass
else:
iseq.msa = themsa
# DEEP META PSICOV RUN
ppaths.mdir(dmpdir)
if skipexec is False:
logger.info(
'Generating contact prediction lists via DeepMetaPSICOV...')
for i, iseq in seq.imer.items():
sfile = fcropseq[i] if cropping is True else fseq[i]
afile = fcropmsa[i] if cropping is True else fmsa[i]
nsfile = os.path.join(dmpdir, os.path.basename(sfile))
if sfile != nsfile:
copyfile(sfile, nsfile)
logger.info(pcl.running('DeepMetaPSICOV'))
itime = datetime.datetime.now()
psd.rundmp(nsfile, afile, dmpdir)
logger.info(pcl.running('DeepMetaPSICOV', done=itime))
# INTERFACE GENERATION, PISA
ppaths.mdir(pisadir)
if skipexec is False:
logger.info('Generating interface files via PISA...')
sfile = fcropstr if cropping is True else fstr
logger.info(pcl.running('PISA'))
itime = datetime.datetime.now()
iflist = psp.runpisa(sfile, pisadir, sessionid=pdbid)
logger.info(pcl.running('PISA', done=itime))
endmsg = pcl.ok(starttime, command=__script__)
logger.info(endmsg)
return
def main():
starttime = time.time()
parser = create_argument_parser()
args = parser.parse_args()
global logger
logger = pcl.pisacov_logger(level="info")
welcomemsg, starttime = pcl.welcome(command=__script__)
logger.info(welcomemsg)
# PARSE CONFIGURATION FILE:
invals = pco._initialise_inputs()
invals['INSEQ'] = None
invals['INIFS'] = None
invals['OUTROOT'] = None
invals['OUTCSVPATH'] = None
# READ INPUT ARGUMENTS
invals['INSEQ'] == ppaths.check_path(args.seqpath[0], 'file')
invals['INIFS'] = []
args.remove_insertions = False
for fp in args.dimers:
if '*' in fp:
invals['INIFS'] += ppaths.check_wildcard(fp)
else:
invals['INIFS'].append(ppaths.check_path(fp, 'file'))
invals['INIFS'] = list(dict.fromkeys(invals['INIFS']))
if args.hhblits_arguments is not None:
invals['HHBLITS_PARAMETERS'] = pco._check_hhparams(
args.hhblits_arguments)
else:
pass
if args.skip_conpred is True:
skipexec = True
if args.hhblits_arguments is not None:
logger.info('HHblits parameters given bypassed by --skip_conpred')
else:
skipexec = False
if args.outdir is None:
invals['OUTROOT'] = ppaths.check_path(os.path.dirname(invals['INSEQ']))
else:
invals['OUTROOT'] = ppaths.check_path(os.path.join(args.outdir[0], ''))
ppaths.mdir(invals['OUTROOT'])
if args.collection_file is None:
invals['OUTCSVPATH'] = ppaths.check_path(
os.path.join(invals['OUTROOT'],
("evcovsignal" + os.extsep + "full" + os.extsep +
"pisacov" + os.extsep + "csv")))
else:
invals['OUTCSVPATH'] = ppaths.check_path(args.collection_file[0])
if os.path.isfile(invals['OUTCSVPATH']) is False:
pic.csvheader(invals['OUTCSVPATH'], cropped=False)
# Define formats used
sources = pco._sources()
# Parse sequence and structure files
logger.info('Parsing sequence file...')
seqs = cps.parseseqfile(invals['INSEQ'])
if len(seqs) == 1:
if len(seqs) == 1:
for key in seqs:
pdbid = key.lower()
else:
raise Exception('More than one pdbid in sequence set.')
seq = seqs[pdbid]
outpdbdir = os.path.join(invals['OUTROOT'], pdbid, "")
# RENUMBERING
fseq = {}
fmsa = {}
if skipexec is False:
if invals['INIFS'] is not None:
logger.info('Renumbering interfaces provided ' +
'according to position in sequence.')
for path in invals['INIFS']:
instrc = os.path.join(invals['OUTROOT'], pdbid,
os.path.basename(path))
logger.info(pcl.running('CROPS-renumber'))
itime = datetime.datetime.now()
psc.renumcrops(invals['INSEQ'], path, invals['OUTROOT'])
copyfile(path, instrc)
logger.info(pcl.running('CROPS-renumber', done=itime))
ppaths.mdir(outpdbdir)
for i, iseq in seq.imer.items():
fiseq = pdbid + '_' + i + os.extsep + 'fasta'
fseq[i] = os.path.join(invals['OUTROOT'], pdbid, fiseq)
fiseq = pdbid + '_' + i + os.extsep + 'msa' + os.extsep + 'aln'
fmsa[i] = os.path.join(invals['OUTROOT'], pdbid, 'hhblits', fiseq)
if skipexec is False:
iseq.dump(fseq[i])
# EXECUTION OF EXTERNAL PROGRAMS
hhdir = os.path.join(invals['OUTROOT'], pdbid, 'hhblits', '')
dmpdir = os.path.join(invals['OUTROOT'], pdbid, 'dmp', '')
fstr = []
for file in invals['INIFS']:
fstr.append(
os.path.join(
invals['OUTROOT'],
(os.path.splitext(os.path.basename(file))[0] + os.extsep +
'crops' + os.extsep + 'seq' + os.extsep + 'pdb')))
if skipexec is False:
# MSA GENERATOR
ppaths.mdir(hhdir)
if invals['HHBLITS_PARAMETERS'] == ['3', '0.001', 'inf', '50', '99']:
logger.info(
'Generating Multiple Sequence Alignment using DeepMetaPSICOV default parameters... [AS RECOMMENDED]'
)
elif invals['HHBLITS_PARAMETERS'] == ['2', '0.001', '1000', '0', '90']:
logger.info(
'Generating Multiple Sequence Alignment using HHBlits default parameters...'
)
else:
logger.info(
'Generating Multiple Sequence Alignment using user-custom parameters...'
)
for i, iseq in seq.imer.items():
sfile = fseq[i]
afile = fmsa[i]
logger.info(pcl.running('HHBlits'))
itime = datetime.datetime.now()
themsa = psm.runhhblits(sfile, invals['HHBLITS_PARAMETERS'], hhdir)
logger.info(pcl.running('HHBlits', done=itime))
iseq.msa = themsa
# DEEP META PSICOV RUN
logger.info(
'Generating contact prediction lists via DeepMetaPSICOV...')
ppaths.mdir(dmpdir)
for i, iseq in seq.imer.items():
sfile = fseq[i]
afile = fmsa[i]
nsfile = os.path.join(dmpdir, os.path.basename(sfile))
if sfile != nsfile:
copyfile(sfile, nsfile)
logger.info(pcl.running('DeepMetaPSICOV'))
itime = datetime.datetime.now()
psd.rundmp(nsfile, afile, dmpdir)
logger.info(pcl.running('DeepMetaPSICOV', done=itime))
# GENERATE INTERFACE LIST
iflist = []
for filepath in fstr:
ifname = os.path.splitext(os.path.basename(filepath))[0]
iflist.append(pci.interface(name=ifname))
# CONTACT ANALYSIS AND MATCH
logger.info('Opening output csv files...')
resultdir = os.path.join(invals['OUTROOT'], pdbid, 'pisacov', '')
ppaths.mdir(resultdir)
csvfile = os.path.join(
resultdir, (pdbid + os.extsep + "evcovsignal" + os.extsep + "full" +
os.extsep + "pisacov" + os.extsep + "csv"))
pic.csvheader(csvfile, cropped=False, pisascore=False)
logger.info('Parsing sequence files...')
for i, fpath in fseq.items():
seq.imer[i].seqs['conkit'] = ckio.read(fpath, 'fasta')[0]
seq.imer[i].biotype = csq.guess_type(seq.imer[i].seqs['mainseq'])
logger.info('Parsing contact predictions lists...')
conpred = {}
matches = []
for s in seq.imer:
if s not in conpred:
conpred[s] = {}
for source, attribs in sources.items():
fc = os.path.splitext(os.path.basename(fseq[s]))[0]
fc += attribs[1]
confile = os.path.join(invals['OUTROOT'], pdbid, attribs[0], fc)
conpred[s][source] = ckio.read(confile, attribs[2])[0]
logger.info('Parsing crystal structure contacts...')
for i in range(len(iflist)):
inputmap = ckio.read(fstr[i], 'pdb')
if len(inputmap) == 4:
chnames = list(iflist[i].chains.keys())
chtypes = list(iflist[i].chains.values())
if (seq.whatseq(chnames[0]) != seq.whatseq(chnames[1])
or (chtypes[0] != 'Protein' or chtypes[1] != 'Protein')):
if chtypes[0] != "Protein" or chtypes[1] != "Protein":
logger.info(
'Interface ' + str(i) +
' is not a Protein-Protein interface. Ignoring.')
else:
logger.info('Interface ' + str(i) +
' is not a homodimer. Ignoring.')
iflist[i].structure = None
matches.append(None)
continue
s = seq.whatseq(chnames[0])
try:
iflist[i].structure = []
for m in range(len(inputmap)):
iflist[i].structure.append(inputmap[m].as_contactmap())
iflist[i].structure[m].id = inputmap[m].id
except Exception:
for m in range(len(inputmap)):
iflist[i].structure.append(inputmap[m]) # ConKit LEGACY.
matches.append({})
for source, attribs in sources.items():
matches[i][source] = pcc.contact_atlas(
name=pdbid + '_' + str(s),
conpredmap=conpred[s][source],
strmap=iflist[i].structure,
sequence=seq.imer[s],
removeintra=True)
else:
iflist[i].structure = None
matches.append(None)
continue
logger.info('Computing results and writing them to file...')
for i in range(len(iflist)):
if matches[i] is None:
continue
results = [pdbid, str(i + 1)]
results.append(matches[i]['psicov'].chain1)
results.append(matches[i]['psicov'].chain2)
sid = seq.whatseq(matches[i]['psicov'].chain1)
results.append(str(sid))
results.append(str(seq.imer[sid].length()))
results.append(str(seq.imer[sid].cropmsa.meff))
results.append(str(seq.imer[sid].ncrops()))
results.append(str(seq.imer[sid].full_length()))
for source, attribs in sources.items():
appresults = pcs.list_scores(matches[i][source], tag=source)
results += appresults
pic.lineout(results, csvfile)
pic.lineout(results, invals['OUTCSVPATH'])
endmsg = pcl.ok(starttime, command=__script__)
logger.info(endmsg)
return
def _check_input(val, key):
compmsg = {
'SIFTS_PATH': "SIFTS_PATH file not found.",
'PISA_PATH': "'PISA_PATH file not found.",
'HHBLITS_PATH': "HHBLITS_PATH file not found.",
'HHBLITS_DATABASE_NAME': "HHBLITS_DATABASE_DIR should be a string.",
'HHBLITS_DATABASE_DIR': "HHBLITS_DATABASE_DIR directory not found.",
'DMP_PATH': "DMP_PATH file not found.",
'HHBLITS_PARAMETERS':
"Format should be [int, float, int or 'inf', int, int].",
'UNICLUST_FASTA_PATH': "UNICLUST_FASTA_PATH file not found.",
'NEIGHBOURS_MINDISTANCE':
"NEIGHBOURS_MINDISTANCE should be an integer.",
'REMOVE_INTRA_CONTACTS': "REMOVE_INTRA_CONTACTS should be a boolean."
}
errormsg = compmsg[
key] + ' Please, update the configuration file using pisaconf.'
if (key == 'SIFTS_PATH' or key == 'PISA_PATH' or key == 'HHBLITS_PATH'
or key == 'DMP_PATH'):
try:
val = check_path(val, 'file')
except Exception:
logging.critical(errormsg)
raise NameError()
if (key == 'UNICLUST_FASTA_PATH'):
if val == "" or val is None:
val = None
else:
try:
val = check_path(val, 'file')
except Exception:
logging.critical(errormsg)
raise NameError()
elif (key == 'HHBLITS_DATABASE_DIR'):
try:
val = check_path(val, 'dir')
except Exception:
logging.critical(errormsg)
raise NameError()
elif (key == 'HHBLITS_DATABASE_NAME'):
if isinstance(val, str) is False:
try:
val = str(val)
except Exception:
logging.critical(errormsg)
raise ValueError()
else:
pass
elif (key == 'HHBLITS_PARAMETERS'):
if val is None or val == "":
val = _default_values(key)
else:
if isinstance(val, list) is False:
logging.critical('HHBLITS_PARAMETERS is not a python list.')
raise ValueError()
else:
for n in range(len(val)):
if n == 1:
if isinstance(val[n], float) is False:
try:
val[n] = float(val[n])
except Exception:
logging.critical(errormsg)
raise TypeError
else:
if n == 2 and val[n] == 'inf':
pass
else:
if isinstance(val[n], int) is False:
try:
val[n] = int(val[n])
except Exception:
logging.critical(errormsg)
raise TypeError
elif (key == 'NEIGHBOURS_MINDISTANCE'):
if val is None or val == "":
val = _default_values(key)
else:
if isinstance(val[n], int) is False:
try:
val = int(val)
except Exception:
logging.critical(errormsg)
raise TypeError
elif (key == 'REMOVE_INTRA_CONTACTS'):
if val is None:
val = _default_values(key)
elif isinstance(val, str) is True:
if val == "":
val = _default_values(key)
elif val.lower() == "true":
val = True
elif val.lower() == "false":
val = False
else:
logging.critical(errormsg)
raise TypeError
elif isinstance(val, bool) is True:
pass
else:
logging.critical(errormsg)
raise TypeError
return val
def main():
parser = create_argument_parser()
args = parser.parse_args()
global logger
logger = pcl.pisacov_logger(level="info")
welcomemsg, starttime = pcl.welcome(command=__script__)
logger.info(welcomemsg)
# PARSE CONFIGURATION FILE:
invals = pco._initialise_inputs()
invals['INSEQ'] = None
invals['INSTR'] = None
invals['ALTDB'] = None
invals['OUTROOT'] = None
invals['OUTCSVPATH'] = None
invals['UPTHRESHOLD'] = None
# READ INPUT ARGUMENTS
invals['INSEQ'] = ppaths.check_path(args.seqpath[0], 'file')
invals['INSTR'] = ppaths.check_path(args.crystalpath[0], 'file')
if args.hhblits_arguments is not None:
invals['HHBLITS_PARAMETERS'] = pco._check_hhparams(
args.hhblits_arguments)
else:
pass
if args.uniprot_threshold is not None:
try:
invals['UPTHRESHOLD'] = float(args.uniprot_threshold[0])
except ValueError:
logger.critical('Uniprot threshold given not valid.')
if invals['UNICLUST_FASTA_PATH'] is None:
invals['UNICLUST_FASTA_PATH'] = pco._uniurl
else:
pass
if args.skip_conpred is True:
skipexec = True
if (args.hhblits_arguments is not None
or args.uniprot_threshold is not None):
logger.info(
'HHblits, UniProt threshold parameters given bypassed by --skip_conpred'
)
else:
skipexec = False
cropping = args.remove_insertions
scoring = [cropping, not cropping]
if args.outdir is None:
invals['OUTROOT'] = ppaths.check_path(os.path.dirname(invals['INSEQ']))
else:
invals['OUTROOT'] = ppaths.check_path(os.path.join(args.outdir[0], ''))
ppaths.mdir(invals['OUTROOT'])
invals['OUTCSVPATH'] = []
if args.collection_file is None:
invals['OUTCSVPATH'].append(
ppaths.check_path(
os.path.join(invals['OUTROOT'],
("evcovsignal" + os.extsep + "cropped" +
os.extsep + "pisacov" + os.extsep + "csv"))))
invals['OUTCSVPATH'].append(
ppaths.check_path(
os.path.join(invals['OUTROOT'],
("evcovsignal" + os.extsep + "full" + os.extsep +
"pisacov" + os.extsep + "csv"))))
else:
if cropping is True:
invals['OUTCSVPATH'].append(
ppaths.check_path(args.collection_file[0]))
invals['OUTCSVPATH'].append(
ppaths.check_path(
os.path.splitext(args.collection_file[0])[0] + os.extsep +
'full' + os.extsep +
os.path.splitext(args.collection_file[0])[1]))
else:
invals['OUTCSVPATH'].append(None)
invals['OUTCSVPATH'].append(
ppaths.check_path(args.collection_file[0]))
if args.plot_formats is None:
plotformats = {'png'}
else:
plotformats = set()
for element in args.plot_formats:
if element.lower() in {'png', 'eps', 'dat'}:
plotformats.add(element.lower())
# Define formats used
sources = pco._sources()
# Parse sequence and structure files
logger.info('Parsing sequence file...')
# seqs = cps.parseseqfile(invals['INSEQ'])
seqs = pio.read(invals['INSEQ'], 'fasta')
logger.info('Parsing structure file...')
# strs, filestrs = cps.parsestrfile(invals['INSTR'])
strs, filestrs = pio.read(invals['INSTR'], 'pdb')
if len(seqs) == 1 or len(strs) == 1:
if len(seqs) == 1:
for key in seqs:
pdbid = key
elif len(seqs) > 1 and len(strs) == 1:
for key in strs:
for key2 in seqs:
if key.upper() == key2.upper():
pdbid = key.upper()
else:
if key2.upper() in key.upper():
pdbid = key2.upper()
else:
raise Exception(
'More than one pdbid in sequence and/or structure set.')
seq = seqs[pdbid]
#structure = strs[pdbid]
# CROPPING AND RENUMBERING
outpdbdir = os.path.join(invals['OUTROOT'], pdbid, "")
instrc = os.path.join(invals['OUTROOT'], pdbid,
os.path.basename(invals['INSTR']))
fseq = {}
fmsa = {}
if skipexec is False:
if cropping is True:
logger.info('Cropping and renumbering sequences, ' +
'structures according to SIFTS database.')
logger.info(pcl.running('CROPS-cropstr'))
itime = datetime.datetime.now()
psc.runcrops(invals['INSEQ'], invals['INSTR'],
invals['SIFTS_PATH'], invals['UPTHRESHOLD'],
invals['UNICLUST_FASTA_PATH'], invals['OUTROOT'])
logger.info(pcl.running('CROPS-cropstr', done=itime))
else:
logger.info('Renumbering structure ' +
'according to position in sequence.')
logger.info(pcl.running('CROPS-renumber'))
itime = datetime.datetime.now()
psc.renumcrops(invals['INSEQ'], invals['INSTR'], invals['OUTROOT'])
logger.info(pcl.running('CROPS-renumber', done=itime))
ppaths.mdir(outpdbdir)
copyfile(invals['INSTR'], instrc)
for i, iseq in seq.imer.items():
fiseq = pdbid + '_' + i + '.fasta'
fseq[i] = os.path.join(invals['OUTROOT'], pdbid, fiseq)
fiseq = pdbid + '_' + i + '.msa.aln'
fmsa[i] = os.path.join(invals['OUTROOT'], pdbid, 'hhblits', fiseq)
if skipexec is False:
iseq.dump(fseq[i])
# Parse cropped sequences and maps
if cropping is True:
amap = {}
fcropseq = {}
fcropmsa = {}
for i, iseq in seq.imer.items():
fprefix = pdbid + '_' + i + '.crops.to_uniprot'
fmap = os.path.join(invals['OUTROOT'], pdbid,
fprefix + os.extsep + 'cropmap')
amap.update(cps.parsemapfile(fmap)[pdbid])
fcropseq[i] = os.path.join(invals['OUTROOT'], pdbid,
fprefix + os.extsep + 'fasta')
fcropmsa[i] = os.path.join(
invals['OUTROOT'], pdbid, 'hhblits',
(fprefix + os.extsep + 'msa' + os.extsep + 'aln'))
seq.set_cropmaps(amap, cropmain=True)
if iseq.ncrops() == 0:
logger.info(' Cropped sequence ' + iseq.oligomer_id + '_' +
iseq.name +
' is identical to the original sequence.')
else:
logger.info(' Cropped sequence ' + iseq.oligomer_id + '_' +
iseq.name + ' is ' + str(iseq.ncrops()) +
' residues ' +
'shorter than the original sequence.')
# EXECUTION OF EXTERNAL PROGRAMS
hhdir = os.path.join(invals['OUTROOT'], pdbid, 'hhblits', '')
dmpdir = os.path.join(invals['OUTROOT'], pdbid, 'dmp', '')
pisadir = os.path.join(invals['OUTROOT'], pdbid, 'pisa', '')
fstr = os.path.join(
invals['OUTROOT'],
(pdbid + os.extsep + 'crops' + os.extsep + 'seq' + os.extsep + 'pdb'))
if cropping:
fcropstr = os.path.join(
invals['OUTROOT'], pdbid,
(pdbid + os.extsep + 'crops' + os.extsep + 'oldids' + os.extsep +
'to_uniprot' + os.path.splitext(invals['INSTR'])[1]))
if skipexec is False:
# MSA GENERATOR
ppaths.mdir(hhdir)
if invals['HHBLITS_PARAMETERS'] == ['3', '0.001', 'inf', '50', '99']:
logger.info(
'Generating Multiple Sequence Alignment using DeepMetaPSICOV default parameters... [AS RECOMMENDED]'
)
elif invals['HHBLITS_PARAMETERS'] == ['2', '0.001', '1000', '0', '90']:
logger.info(
'Generating Multiple Sequence Alignment using HHBlits default parameters...'
)
else:
logger.info(
'Generating Multiple Sequence Alignment using user-custom parameters...'
)
for i, iseq in seq.imer.items():
sfile = fcropseq[i] if cropping is True else fseq[i]
afile = fcropmsa[i] if cropping is True else fmsa[i]
logger.info(pcl.running('HHBlits'))
itime = datetime.datetime.now()
themsa = psm.runhhblits(sfile, invals['HHBLITS_PARAMETERS'], hhdir)
logger.info(pcl.running('HHBlits', done=itime))
if cropping is True:
iseq.cropmsa = themsa
if iseq.ncrops() == 0:
iseq.msa = iseq.cropmsa
continue
else:
pass
else:
iseq.msa = themsa
# DEEP META PSICOV RUN
ppaths.mdir(dmpdir)
if skipexec is False:
logger.info(
'Generating contact prediction lists via DeepMetaPSICOV...')
for i, iseq in seq.imer.items():
sfile = fcropseq[i] if cropping is True else fseq[i]
afile = fcropmsa[i] if cropping is True else fmsa[i]
nsfile = os.path.join(dmpdir, os.path.basename(sfile))
if sfile != nsfile:
copyfile(sfile, nsfile)
logger.info(pcl.running('DeepMetaPSICOV'))
itime = datetime.datetime.now()
psd.rundmp(nsfile, afile, dmpdir)
logger.info(pcl.running('DeepMetaPSICOV', done=itime))
# INTERFACE GENERATION, PISA
ppaths.mdir(pisadir)
if skipexec is False:
logger.info('Generating interface files via PISA...')
sfile = fcropstr if cropping is True else fstr
logger.info(pcl.running('PISA'))
itime = datetime.datetime.now()
iflist = psp.runpisa(sfile, pisadir, sessionid=pdbid)
logger.info(pcl.running('PISA', done=itime))
# READ DATA IF SKIPEXEC USED:
if skipexec is True:
logger.info('Parsing already generated files...')
for i, iseq in seq.imer.items():
sfile = fcropstr if cropping is True else fstr
afile = fcropmsa[i] if cropping is True else fmsa[i]
if cropping is True:
# iseq.cropmsa = ckio.read(afile, 'jones')
iseq.cropmsa = pio.read(afile, 'jones')
if iseq.ncrops() == 0:
scoring[1] = True
# iseq.msa = ckio.read(afile, 'jones')
iseq.msa = ckio.read(afile, 'jones')
else:
# iseq.msa = ckio.read(afile, 'jones')
iseq.msa = pio.read(afile, 'jones')
ixml = os.path.join(pisadir,
(os.path.splitext(os.path.basename(sfile))[0] +
os.extsep + 'interface' + os.extsep + 'xml'))
axml = os.path.join(pisadir,
(os.path.splitext(os.path.basename(sfile))[0] +
os.extsep + 'assembly' + os.extsep + 'xml'))
iflist = pci.parse_interface_xml(ixml, axml)
# CONTACT ANALYSIS AND MATCH
logger.info('Opening output csv files...')
resultdir = os.path.join(invals['OUTROOT'], pdbid, 'pisacov', '')
ppaths.mdir(resultdir)
csvfile = []
csvfile.append(
os.path.join(resultdir,
(pdbid + os.extsep + "evcovsignal" + os.extsep +
"cropped" + os.extsep + "pisacov" + os.extsep + "csv")))
csvfile.append(
os.path.join(resultdir,
(pdbid + os.extsep + "evcovsignal" + os.extsep + "full" +
os.extsep + "pisacov" + os.extsep + "csv")))
for n in range(2):
if scoring[n] is True:
cpd = True if cropping else False
pic.csvheader(csvfile[n], cropped=cpd, pisascore=True)
if invals['OUTCSVPATH'][n] is not None:
if os.path.isfile(invals['OUTCSVPATH'][n]) is False:
pic.csvheader(invals['OUTCSVPATH'][n],
cropped=cpd,
pisascore=True)
logger.info('Parsing sequence files...')
for i, fpath in fseq.items():
# seq.imer[i].seqs['conkit'] = ckio.read(fpath, 'fasta')[0]
seq.imer[i].seqs['conkit'] = pio.read(fpath, 'fasta', ck=True)[0]
logger.info('Parsing contact predictions lists...')
conpred = {}
matches = []
for s in seq.imer:
if s not in conpred:
conpred[s] = {}
fs = fcropseq[s] if cropping else fseq[s]
for source, attribs in sources.items():
fc = os.path.splitext(os.path.basename(fs))[0]
fc += os.extsep + attribs[1]
confile = os.path.join(dmpdir, fc)
# conpred[s][source] = ckio.read(confile, attribs[2])[0]
conpred[s][source] = pio.read(confile, attribs[2], ck=True)[0]
logger.info('Parsing crystal structure contacts...')
for i in range(len(iflist)):
logger.info(os.linesep + str(iflist[i]))
fs = fcropstr if cropping else fstr
fs = (os.path.splitext(os.path.basename(fs))[0] + os.extsep +
"interface" + os.extsep + str(i + 1) + os.extsep + "pdb")
spath = os.path.join(pisadir, fs)
# inputmap = ckio.read(spath, 'pdb')
inputmap = pio.read(spath, 'pdb', ck=True)
if len(inputmap) == 4:
chnames = [
iflist[i].chains[0].crystal_id, iflist[i].chains[1].crystal_id
]
iflist[i].chains[0].seq_id = seq.whatseq(chnames[0])
iflist[i].chains[1].seq_id = seq.whatseq(chnames[1])
chseqs = [iflist[i].chains[0].seq_id, iflist[i].chains[1].seq_id]
logger.info(iflist[i].chains)
chtypes = [iflist[i].chains[0].type, iflist[i].chains[1].type]
if (chseqs[0] != chseqs[1]
or (chtypes[0] != 'Protein' or chtypes[1] != 'Protein')):
if chtypes[0] != "Protein" or chtypes[1] != "Protein":
logger.info(
'Interface ' + str(i) +
' is not a Protein-Protein interface. Ignoring.')
else:
logger.info('Interface ' + str(i) +
' is not a homodimer. Ignoring.')
iflist[i].structure = None
matches.append(None)
continue
s = chseqs[0]
try:
iflist[i].structure = []
for m in range(len(inputmap)):
iflist[i].structure.append(inputmap[m].as_contactmap())
iflist[i].structure[m].id = inputmap[m].id
except Exception:
logger.warning('Contact Maps obtained from a legacy ConKit ' +
'version with no Distograms implemented.')
for m in range(len(inputmap)):
iflist[i].structure.append(inputmap[m]) # ConKit LEGACY.
#fs = fcropstr if cropping else fstr
#fs = (os.path.splitext(os.path.basename(fs))[0] +
# os.extsep + "interface" + os.extsep + str(i+1) + os.extsep + "con")
#spath = os.path.join(pisadir, fs)
#pio.write(spath, 'psicov', indata=iflist[i].structure[1])
#iflist[i].contactmap = pio.read(spath, 'array')
iflist[i].contactmap = iflist[i].structure[1].deepcopy()
matches.append({})
for source, attribs in sources.items():
matches[i][source] = pcc.contact_atlas(
name=pdbid + '_' + str(s),
dimer_interface=iflist[i],
conpredmap=conpred[s][source],
conpredtype=source,
sequence=seq.imer[s])
if cropping is True:
matches[i][source].set_cropmap()
matches[i][source].remove_neighbours(mindist=2)
matches[i][source].set_conpred_seq()
matches[i][source].remove_intra()
matches[i][source].make_match(filterout=attribs[3])
for cmode, cmap in matches[i][source].conkitmatch.items():
if (len(cmap) > 0 and len(
matches[i][source].interface.structure[1]) > 0):
for imtype in plotformats:
if len(matches[i][source].conkitmatch) > 1:
pout = (os.path.splitext(fs)[0] + os.extsep +
'match' + os.extsep + cmode +
os.extsep + source + os.extsep +
'con' + os.extsep + imtype)
else:
pout = (os.path.splitext(fs)[0] + os.extsep +
'match' + os.extsep + source +
os.extsep + 'con' + os.extsep + imtype)
plotpath = os.path.join(
os.path.dirname(csvfile[0]), pout)
matches[i][source].plot_map_alt(plotpath,
mode=cmode,
plot_type=imtype)
else:
iflist[i].structure = None
iflist[i].contactmap = None
matches.append(None)
continue
logger.info(os.linesep + 'Computing results and writing them to file...' +
os.linesep)
for i in range(len(iflist)):
logger.info('Generating Interface ' + str(i + 1) + ' data...')
if matches[i] is None:
continue
results = [pdbid, str(i + 1)]
results.append(iflist[i].chains[0].crystal_id)
results.append(iflist[i].chains[1].crystal_id)
sid = iflist[i].chains[0].seq_id
results.append(str(sid))
results.append(str(seq.imer[sid].length()))
if cropping is True:
results.append(str(seq.imer[sid].cropmsa.meff))
else:
results.append(str(seq.imer[sid].msa.meff))
results.append(str(seq.imer[sid].ncrops()))
results.append(str(seq.imer[sid].full_length()))
results.append(str(seq.imer[sid].msa.meff))
for source, attribs in sources.items():
appresults = pcs.list_scores(matches[i][source], tag=source)
results.extend(appresults)
results.append(str(iflist[i].stable))
for n in range(2):
if scoring[n] is True:
pic.lineout(results, csvfile[n])
pic.lineout(results, invals['OUTCSVPATH'][n])
endmsg = pcl.ok(starttime, command=__script__)
logger.info(endmsg)
return
def main():
parser = create_argument_parser()
args = parser.parse_args()
global logger
logger = pcl.pisacov_logger(level="info")
welcomemsg, starttime = pcl.welcome(command=__script__)
logger.info(welcomemsg)
csvfile = ppaths.check_path(args.scores[0], 'file')
outdir = ppaths.check_path(args.outdir)
ppaths.mdir(outdir)
# Parsing scores
scores = {}
names = None
thraw = {}
with open(csvfile, 'r') as fin:
scoresin = csv.reader(fin)
for entry in scoresin:
if entry[0][0] != "#":
if (names is None or isinstance(names, str) or
(isinstance(names, list) and len(names) != len(entry))):
names = []
for n in range(len(entry)):
names.append('sc_' + str(n + 1))
else:
if thraw == {}:
for name in names[13:-1]:
thraw[name] = []
if entry[0] not in scores:
scores[entry[0]] = {}
if entry[1] not in scores[entry[0]]:
scores[entry[0]][entry[1]] = []
for sc in (entry.split(sep=', ')[13:-1]):
scores[entry[0]][entry[1]].append(float(sc))
if (entry.split(sep=', ')[-1]) == 'True' or '1':
scores[entry[0]][entry[1]].append(True)
elif (entry.split(sep=', ')[-1]) == 'False' or '0':
scores[entry[0]][entry[1]].append(False)
for n in range(len(names)):
thraw[names[n]].append(scores[entry[0]][entry[1]][n])
else:
if entry.split(
sep=', ')[13:] == scores[entry[0]][entry[1]]:
pass
else:
raise ValueError(
'CSV file contains different values for same interface.'
)
else:
names = entry[1:].split(sep=', ')
# Setting thresholds
thr = {}
FPR = {}
TPR = {}
for key, value in thraw.items():
thr[key] = list(set(thraw)).sort()
FPR[key] = []
TPR[key] = []
for t in thr[key]:
FP = 0
TP = 0
FN = 0
TN = 0
for pdbid in scores:
for iface in scores[pdbid]:
stable = scores[pdbid][iface][-1]
for n in range(len(names)):
if scores[pdbid][iface][n] < t:
if stable is True:
FN += 1
else:
TN += 1
else:
if stable is True:
TP += 1
else:
FP += 1
FPR[key].append(FP / (FP + TN))
TPR[key].append(TP / (TP + FN))
fnameout = os.path.join(
outdir,
(key + os.path.splitext(os.path.basename(csvfile))[0] + 'roc.dat'))
with open(fnameout, 'w') as fout:
for n in range(len(FPR[key])):
fout.write(str(FPR[key][n]) + ' ' + str(TPR[key][n]))
endmsg = pcl.ok(starttime, command=__script__)
logger.info(endmsg)
return