def write(infile, ftype, indata, ck=False):
"""
:param infile: Path to input file.
:type infile: str
:param ftype: One of 'psicov', 'ccmpred', 'fasta', 'pdb', 'a3m', 'jones', 'xml'.
:type ftype: str
:param ck: Open alternative conkit version instead of default, defaults to False.
:type ck: bool, optional
:return: Parsed file (and, for 'pdb', list of filenames).
:rtype: One or two of list[str], :class:`~crops.elements.sequences.sequence`, :class:`~conkit.core.sequence.Sequence`,
"""
if (ftype.lower() not in _ftypelist() or
isinstance(ftype, str) is not True):
logging.critical('Specified type not valid.')
raise ValueError
if ck is True and ftype.lower() != 'xml':
output = ckio.write(infile, ftype.lower(), hyerarchy=indata)
else:
if ftype.lower() == 'psicov':
pass
if ftype.lower() == 'ccmpred':
pass
elif ftype.lower() == 'fasta':
output = cps.parseseqfile(infile)
elif ftype.lower() == 'pdb':
output1, output2 = cps.parsestrfile(infile)
return output1, output2
elif ftype.lower() == 'a3m' or 'jones':
output = ckio.read(infile, ftype.lower())
elif ftype.lower() == 'xml':
output = ET.parse(infile)
return output
def msafilesgen(inpath_a3m):
"""
Convert a3m format alignment to "jones" format (.aln) and print msa coverage graph
Parameters
----------
inpath_a3m : str
Multiple Sequence Alignment (.a3m) path
Returns
-------
parsemsa : msa
ConKit parsed MSA.
outpath_aln : str
MSA (jones format) file path.
"""
outpath_aln = os.path.join(os.path.splitext(inpath_a3m)[0] + ".aln")
biotag = os.path.splitext(os.path.splitext(inpath_a3m)[0])[1]
msacoveragefile = pdbid(
) + biotag + ".msa.coverage.png" if biotag == ".bio" else pdbid(
) + ".msa.coverage.png"
msacoveragepath = os.path.join(output_dir(), msacoveragefile)
parsemsa = ckio.read(inpath_a3m, 'a3m')
ckio.write(outpath_aln, 'jones', parsemsa)
ckplot.SequenceCoverageFigure(parsemsa, file_name=msacoveragepath)
return parsemsa, outpath_aln
def _compute_single(args):
decoy, decoy_format, cmap = args
dmap = read(decoy, decoy_format).top_map
matched = cmap.match(dmap)
shortrange = matched.short_range_contacts
mediumrange = matched.medium_range_contacts
longrange = matched.long_range_contacts
sprec, mprec, lprec = float("NaN"), float("NaN"), float("NaN")
if shortrange.ncontacts > 0:
sprec = shortrange.precision
if mediumrange.ncontacts > 0:
mprec = mediumrange.precision
if longrange.ncontacts > 0:
lprec = longrange.precision
return sprec, mprec, lprec
def msafilesgen(inpath_a3m):
"""Convert a3m format alignment to "jones" format (.aln) and print msa coverage graph.
:param inpath_a3m: Multiple Sequence Alignment (.a3m) path
:type inpath_a3m: str
:return: ConKit parsed MSA
:rtype: :obj:`~conkit.core.sequencefile.SequenceFile`
"""
parsedmsa = ckio.read(inpath_a3m, 'a3m')
# Convert to 'jones' format
outpath_aln = os.path.join(os.path.splitext(inpath_a3m)[0], ".aln")
ckio.write(outpath_aln, 'jones', parsedmsa)
# Plot Coverage
msacoveragepath = os.path.join(
os.path.splitext(inpath_a3m)[0], ".coverage.png")
fig = ckplot.SequenceCoverageFigure(parsedmsa)
fig.savefig(msacoveragepath, overwrite=True)
neff = parsedmsa.meff
return neff
def main():
starttime = time.time()
parser = create_argument_parser()
args = parser.parse_args()
global logger
logger = pcl.pisacov_logger(level="info")
logger.info(pcl.welcome())
# READ INPUT ARGUMENTS
if args.initialise is not None:
inseq = pio.check_path(args.initialise[0], 'file')
instr = pio.check_path(args.initialise[1], 'file')
indb = pio.check_path(pio.conf.CSV_CHAIN_PATH, 'file')
skipexec = [False, True] if not args.add_noncropped else [False, False]
scoring = [True, False] if not args.add_noncropped else [True, True]
elif args.skip_conpred is not None:
inseq = pio.check_path(args.skip_conpred[0], 'file')
instr = pio.check_path(args.skip_conpred[1], 'file')
skipexec = [True, True]
scoring = [True, False] if not args.add_noncropped else [True, True]
elif args.skip_default_conpred is not None:
inseq = pio.check_path(args.skip_default_conpred[0], 'file')
instr = pio.check_path(args.skip_default_conpred[1], 'file')
skipexec = [True, True] if not args.add_noncropped else [True, False]
scoring = [True, False] if not args.add_noncropped else [True, True]
if args.outdir is None:
outrootdir = pio.check_path(os.path.dirname(inseq))
else:
outrootdir = pio.check_path(os.path.join(args.outdir[0], ''))
pio.mdir(outrootdir)
if args.collection_file is None:
outcsvfile = pio.check_path(
os.path.join(outrootdir, "pisacov_data.csv"))
else:
outcsvfile = pio.check_path(args.collection_file[0])
try:
pio.check_path(outcsvfile, 'file')
csvexists = True
except:
csvexists = False
if args.uniprot_threshold is not None:
thuprot, dbuprot = pio.check_uniprot(args.uniprot_threshold[0])
else:
thuprot = 0.0
dbuprot = None
if args.hhparams is not None:
hhparameters = pio.check_hhparams(args.hhparams)
else:
try:
hhparameters = pio.check_hhparams(pio.conf.HHBLITS_PARAMETERS)
except:
hhparameters = pio.check_hhparams('dmp')
# Define formats used
sources = pio.paths.sources()
n_sources = len(sources)
# Parse sequence and structure files
logger.info('Parsing sequence file...')
seq = cps.parseseqfile(inseq)
if len(seq) == 1:
for key in seq.keys():
pdbid = key.lower()
if len(seq[key].imer) == 1:
for key2 in seq[key].imer[key2]:
chid = key2
else:
raise Exception('More than one pdbid in sequence set.')
else:
raise Exception('More than one pdbid in sequence set.')
logger.info('Parsing structure file...')
structure = cps.parsestrfile(instr)[0][pdbid]
#logger.info('Parsing SIFTS database file...')
#sifts = cps.import_db(indb, pdb_in=pdbid)
# CROPPING AND RENUMBERING
if not skipexec[0]:
logger.info(
'Cropping and renumbering sequences, structures according to SIFTS database.'
)
psys.crops.runcrops(inseq, instr, indb, thuprot, dbuprot, outrootdir)
outpdbdir = os.path.join(outrootdir, pdbid, "")
pio.mdir(outpdbdir)
inseqc = os.path.join(outpdbdir, os.path.basename(inseq))
instrc = os.path.join(outpdbdir, os.path.basename(instr))
copyfile(inseq, inseqc)
copyfile(instr, instrc)
cmappath = os.path.join(os.path.splitext(inseqc), '.cropmap')
# MSA GENERATOR
cseqpath = os.path.join(outpdbdir, pdbid + '.crops.to_uniprot.fasta')
hhdir = os.path.join(outpdbdir, 'hhblits', '')
pio.mdir(hhdir)
neff = {}
if not skipexec[0] or not skipexec[1]:
if hhparameters == ['3', '0.001', 'inf', '50', '99']:
logger.info(
'Generating Multiple Sequence Alignment using DeepMetaPSICOV default parameters... [AS RECOMMENDED]'
)
elif hhparameters == ['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...'
)
if os.path.isfile(cmappath) and not skipexec[0]:
psys.msagen.runhhblits(cseqpath, hhparameters, hhdir)
cmsaa3mfile = os.path.splitext(
os.path.basename(cseqpath))[0] + ".msa.a3m"
cmsaa3mpath = os.path.join(hhdir, cmsaa3mfile)
neff['cropped'] = psys.msagen.msafilesgen(cmsaa3mpath)
neff['original'] = None
if not skipexec[1]:
logger.info(
' Repeating process for non-default sequence...')
neff['cropped'] = None
else:
logger.info(
' No cropped sequence found, using original sequence instead...'
)
if not os.path.isfile(cmappath) or not skipexec[1]:
psys.msagen.runhhblits(inseq, hhparameters, hhdir)
msaa3mfile = os.path.splitext(
os.path.basename(inseq))[0] + ".msa.a3m"
msaa3mpath = os.path.join(hhdir, msaa3mfile)
neff['original'] = psys.msagen.msafilesgen(msaa3mpath)
# DEEP META PSICOV RUN
if not skipexec[0] or not skipexec[1]:
logger.info(
'Generating contact prediction lists via DeepMetaPSICOV...')
dmpdir = os.path.join(outpdbdir, 'dmp', '')
pio.mdir(dmpdir)
if os.path.isfile(cmappath) and not skipexec[0]:
psys.dmp.rundmp(cseqpath, cmsaa3mpath, dmpdir)
if not skipexec[1]:
logger.info(
' Repeating process for non-default sequence...')
else:
logger.info(
' No cropped sequence found, using original sequence instead...'
)
if not os.path.isfile(cmappath) or not skipexec[1]:
psys.dmp.rundmp(inseq, msaa3mpath, dmpdir)
# INTERFACE GENERATION, PISA
cstrpath = os.path.join(outpdbdir, pdbid + '.oldids.crops.to_uniprot.pdb')
pisadir = os.path.join(outpdbdir, 'pisa', '')
n_ifaces = {}
if not skipexec[0] or not skipexec[1]:
logger.info('Generating interface files via PISA...')
if os.path.isfile(cmappath) and not skipexec[0]:
n_ifaces['cropped'] = psys.pisa.runpisa(cstrpath, pisadir)
if not skipexec[1]:
logger.info(
' Repeating process for non-default sequence...')
else:
n_ifaces['cropped'] = None
logger.info(
' No cropped sequence found, using original sequence instead...'
)
if not os.path.isfile(cmappath) or not skipexec[1]:
n_ifaces['original'] = psys.pisa.runpisa(instr, pisadir)
else:
n_ifaces['original'] = None
# CONTACT ANALYSIS AND MATCH
resultdir = os.path.join(outpdbdir, 'results', '')
logger.info('Opening output csv files...')
pdbcsvfile = os.path.join(resultdir, pdbid + ".evcovsignal.csv")
pio.outcsv.csvheader(pdbcsvfile)
if not csvexists:
pio.outcsv.csvheader(outcsvfile)
logger.info('Parsing contact predictions lists...')
ckseq = ckio.read(inseq, 'fasta')
conpred = {}
for source, attribs in sources.items():
for mode in ['cropped', 'original']:
seqfile = cseqpath if mode == 'cropped' else inseq
confile = (os.path.splitext(os.path.basename(seqfile))[0] +
attribs[1])
conpath = os.path.join(outpdbdir, attribs[0], confile)
cropmapping = cps.parsemapfile(cmappath)
if os.path.isfile(conpath):
conpred[mode][source] = ckio.read(conpath, attribs[2])[0]
for contact in conpred[mode][source]:
contact.res1_seq = cropmapping[pdbid][chid]['cropbackmap'][
contact.res1_seq]
contact.res2_seq = cropmapping[pdbid][chid]['cropbackmap'][
contact.res2_seq]
else:
conpred[mode][source] = None
# NOT SURE IT IS WORKING WITH SEVERAL CHAINS OF SAME SEQUENCE. CHECK EVERYTHING.
logger.info(' Parsing PISA interfaces...')
interfaces = {}
for mode in ['cropped', 'original']:
if n_ifaces[mode] is not None:
interfaces[mode] = []
for i in range(int(n_ifaces[mode])):
strfile = cstrpath if mode == 'cropped' else instr
pdbfilei = os.path.splitext(strfile)[0] + ".interface." + str(
i + 1) + ".pdb"
interfaces[mode].append(ckio.read(pdbfilei, 'pdb'))
else:
interfaces[mode] = None
# OUTPUT
# CODE TO PRINT NON-REPEATED LINES
# import csv
# rows = csv.reader(open("file.csv", "rb"))
# newrows = []
# for row in rows:
# if row not in newrows:
# newrows.append(row)
# writer = csv.writer(open("file.csv", "wb"))
# writer.writerows(newrows)
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 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