def list_new_files(pdb1_archive, assession_log, verbosity):
'''
Taps into the pdb1 local repository and checks if there are new files there
which should be assessed by curate_homoDB function. It creates a list with
files from the pdb1 database that are newer than the ones last assessed
(registered in the dat file); which means that it takes the previous
assession log as an input in the form of a dictionary where keys are files
and the values correspond to the last assession time.
Called by: curate_homoDB()
'''
new_files = []
pctools.printv('Assessing files in PDB1 archive...', verbosity)
assert os.path.isdir(pdb1_archive), clrs[
'r'] + '\n\n Not able to find PDB archive.\n\n Does "' + pdb1_archive + '" exist?' + clrs[
'n']
pdbfiles = [
os.path.join(dp, f) for dp, dn, filenames in os.walk(pdb1_archive)
for f in filenames if f.endswith(".pdb1.gz")
]
for f in pdbfiles:
filename = f.split('/')[-1]
mod_date = os.path.getctime(f)
if filename not in assession_log or mod_date > float(
assession_log[filename]):
pctools.printv(
clrs['y'] + f + ' should be assessed' + clrs['n'] + '...\n',
verbosity)
new_files.append(f)
return new_files
def update_seqres(verbosity):
'''
Runs wget to update the local seqres database, decompresses it and runs
makeblastdb.
Called by: update_databases()
'''
seqres_dir = os.path.join(choirdb, 'seqres')
if not os.path.isdir(seqres_dir):
os.mkdir(seqres_dir)
seqres_txt = os.path.join(seqres_dir, 'pdb_seqres.txt')
seqres_fasta = os.path.join(seqres_dir, 'seqres.fasta')
pctools.printv('Fetching pdb_seqres.txt...', verbosity)
attempt = 0
while attempt < 3:
try:
wgetout = subprocess.check_output([
'wget', '-m', '-r', '-nH', '--cut-dirs=3', '--user=anonymous',
seqres_ftp, '-P', seqres_dir
],
stderr=subprocess.STDOUT)
break
except:
attempt += 1
if attempt < 3:
print('Attempt ' + str(attempt) + ' failed, trying again.')
if attempt == 3:
print(
'Failed to download seqres in 3 attempts. Try again later.'
)
no_wget = 'seqres.txt.gz’ -- not retrieving'
if no_wget not in wgetout.decode(
'UTF-8') or not os.path.isfile(seqres_txt):
pctools.printv('Decompressing pdb_seqres.txt...', verbosity)
with gzip.open(seqres_txt + '.gz',
'rb') as fin, open(seqres_fasta, 'wb') as fout:
shutil.copyfileobj(fin, fout)
if no_wget not in wgetout.decode(
'UTF-8') or not os.path.isfile(seqres_fasta + '.pal'):
subprocess.run([
makeblastdb_exe, '-in', seqres_fasta, '-parse_seqids', '-dbtype',
'prot', '-blastdb_version', '5', '-out', seqres
])
def update_uniref(verbosity):
'''
Runs wget to update the local uniref50 database, decompresses it and runs
makeblastdb.
Called by: update_databases()
'''
uniref50_fasta = os.path.join(choirdb, 'uniref50/uniref50.fasta')
pctools.printv('Fetching uniref50.fasta...', verbosity)
attempt = 0
while attempt < 3:
try:
wgetout = subprocess.check_output([
'wget', '-m', '-r', '-nH', '--cut-dirs=4', '--user=anonymous',
uniref50_ftp, '-P', choirdb
],
stderr=subprocess.STDOUT)
break
except:
attempt += 1
if attempt < 3:
print('Attempt ' + str(attempt) + ' failed, trying again.')
if attempt == 3:
print(
'Failed to download UniRef50 in 3 attempts. Try again later.'
)
no_wget = 'uniref50.fasta.gz’ -- not retrieving'
if no_wget not in wgetout.decode(
'UTF-8') or not os.path.isfile(uniref50_fasta):
pctools.printv('Decompressing uniref50.fasta...', verbosity)
with gzip.open(uniref50_fasta + '.gz',
'rb') as fin, open(uniref50_fasta, 'wb') as fout:
shutil.copyfileobj(fin, fout)
if no_wget not in wgetout.decode(
'UTF-8') or not os.path.isfile(uniref50_fasta + '.pal'):
subprocess.run([
makeblastdb_exe, '-in', uniref50_fasta, '-parse_seqids', '-dbtype',
'prot', '-out', uniref50
])
def score_pairwise(seq1, seq2, matrix, gap_s, gap_e):
score = 0
gap = False
ipos = 0
fpos = 30
nwindows = -(-len(seq1) // 30)
pctools.printv('Number of 30-residue segments: ' + str(nwindows),
g_args.verbosity)
wscores = []
for window in range(nwindows):
wscore = 0
if fpos > len(seq1):
fpos = len(seq1)
pctools.printv(
str(ipos + 1) + ' ' + seq1[ipos:fpos] + ' ' + str(fpos),
g_args.verbosity)
pctools.printv(
str(ipos + 1) + ' ' + seq2[ipos:fpos] + ' ' + str(fpos),
g_args.verbosity)
for i in range(len(seq1))[ipos:fpos]:
pair = (seq1[i], seq2[i])
if not gap:
if pair == ('-', '-'):
score += 4
wscore += 4
elif '-' in pair:
gap = True
score += gap_s
wscore += gap_s
else:
score += score_match(pair, matrix)
wscore += score_match(pair, matrix)
else:
if '-' not in pair:
gap = False
score += score_match(pair, matrix)
wscore += score_match(pair, matrix)
else:
score += gap_e
wscore += gap_e
ipos += 30
fpos += 30
pctools.printv('Segment score: ' + str(wscore), g_args.verbosity)
wscores.append(wscore)
return score, wscores
def collect_fasta(verbosity):
'''
Fetches fasta files in the pdb_homo_archive and creates a single fasta file
within a "sequences" folder. For that, it checks the identity among the
chains in the original fasta and only keeps track of the unique chains, i.e.
less than 99% identity to the other chains. This file is later use to make
the blast database.
Called by: update_databases()
'''
fastafiles = [
os.path.join(dp, f) for dp, dn, filenames in os.walk(pdb_homo_archive)
for f in filenames if f.endswith(".fasta")
]
seqdir = os.path.join(pdb_homo_archive, 'sequences')
if not os.path.isdir(seqdir):
os.mkdir(seqdir)
largepdb_collected_fasta = os.path.join(seqdir,
'largepdb_collected.fastas')
with open(largepdb_collected_fasta, 'w+'):
pass
homo_collected_fasta = os.path.join(seqdir, 'homo_collected.fastas')
with open(homo_collected_fasta, 'w+'):
pass
mono_collected_fasta = os.path.join(seqdir, 'mono_collected.fastas')
with open(mono_collected_fasta, 'w+'):
pass
hetero_collected_fasta = os.path.join(seqdir, 'hetero_collected.fastas')
with open(hetero_collected_fasta, 'w+'):
pass
for fasta in pg(fastafiles, widgets=widgets):
pctools.printv('Assessing ' + clrs['y'] + fasta + clrs['n'] + '...',
verbosity)
contents = open(fasta, 'r').read()
contentlines = contents.split('>')
nchains = str(len(re.findall('>', contents)))
pctools.printv(
'With ' + clrs['y'] + nchains + clrs['n'] +
' chains to be assessed\n', verbosity)
uniques = []
for entry in contentlines:
if entry:
splitentry = entry.split('\n', 1)
pdbch = splitentry[0]
seq = splitentry[1].replace('\n', '')
if uniques:
percent_ids = []
for unique in uniques:
alignment = parasail.sg_stats_striped_16(
seq, unique[1], 10, 1, parasail.blosum62)
if alignment.length == 0:
percent_ids.append(0)
else:
percent_ids.append(
(alignment.matches) / alignment.length * 100)
if all(percent_id <= 99 for percent_id in percent_ids):
uniques.append([pdbch, seq])
else:
uniques.append([pdbch, seq])
if '/largepdb_sequences/' in fasta:
with open(largepdb_collected_fasta, 'a') as f:
for unique in uniques:
wrapped_seq = "\n".join(tw.wrap(unique[1]))
fasta_entry = '>' + unique[0] + '\n' + wrapped_seq + '\n\n'
f.write(fasta_entry)
elif '/mono_sequences/' in fasta:
with open(mono_collected_fasta, 'a') as f:
for unique in uniques:
wrapped_seq = "\n".join(tw.wrap(unique[1]))
fasta_entry = '>' + unique[0] + '\n' + wrapped_seq + '\n\n'
f.write(fasta_entry)
elif '/hetero_sequences/' in fasta:
with open(hetero_collected_fasta, 'a') as f:
for unique in uniques:
wrapped_seq = "\n".join(tw.wrap(unique[1]))
fasta_entry = '>' + unique[0] + '\n' + wrapped_seq + '\n\n'
f.write(fasta_entry)
elif '/homo_sequences/' in fasta:
with open(homo_collected_fasta, 'a') as f:
for unique in uniques:
wrapped_seq = "\n".join(tw.wrap(unique[1]))
fasta_entry = '>' + unique[0] + '\n' + wrapped_seq + '\n\n'
f.write(fasta_entry)
subprocess.run([
makeblastdb_exe, '-in', largepdb_collected_fasta, '-dbtype', 'prot',
'-out',
os.path.join(seqdir, 'largedb')
])
subprocess.run([
makeblastdb_exe, '-in', mono_collected_fasta, '-dbtype', 'prot',
'-out',
os.path.join(seqdir, 'monodb')
])
subprocess.run([
makeblastdb_exe, '-in', hetero_collected_fasta, '-dbtype', 'prot',
'-out',
os.path.join(seqdir, 'heterodb')
])
subprocess.run([
makeblastdb_exe, '-in', homo_collected_fasta, '-dbtype', 'prot',
'-out',
os.path.join(seqdir, 'homodb')
])
def curate_homoDB(verbosity):
'''
Creates h**o-oligomeric database from a local pdb repsitory.
The divided scheme adopted by RCSB, in which the subdirectories
are the two middle characters in the PDB code, is assumed.
Each database contains three key files: dat, log and fasta.
* homodb.dat contains only the pdb codes contained in the database.
* homodb.log contains summarized relevant information about each entry.
* homodb.fasta contains the sequences of every chain in the database.
Called by: update_databases()
'''
# Create stats folder if does not exist
stats_dir = os.path.join(pdb_homo_archive, 'stats')
if not os.path.isdir(stats_dir):
os.mkdir(stats_dir)
# Compare latest assession with new files
assession_log = read_latest_assession(stats_dir)
new_files = list_new_files(pdb1_archive, assession_log, verbosity)
print(clrs['g'] + str(len(new_files)) + clrs['n'] +
' new structure files were found and will be processed')
now = str(time.strftime("%d-%m-%Y@%H.%M.%S"))
dat_file = os.path.join(stats_dir, now + '-choirdb.dat')
log_file = os.path.join(stats_dir, now + '-choirdb.log')
err_file = os.path.join(stats_dir, now + '-choirdb.err')
if not os.path.isfile(dat_file):
with open(dat_file, 'w+'):
pass
# Write files not to be updated to new dat file
with open(dat_file, 'a') as f:
for i in assession_log:
if i not in new_files:
f.write(i + " " + assession_log[i] + "\n")
# Create log file
if not os.path.isfile(log_file):
with open(log_file, 'w+') as f:
f.write('Code, Chains, Author, Software, Date\n')
# Read Chain correspondences
chain_correspondences_file = os.path.join(stats_dir,
'chain_correspondences.pickle')
if os.path.isfile(chain_correspondences_file):
with open(chain_correspondences_file, 'rb') as p:
chain_correspondences = pickle.load(p)
else:
chain_correspondences = {}
# Main loop that will populate the ProtCHOIR database
for pdb in pg(new_files, widgets=widgets):
filename = pdb.split('/')[-1]
subfolder = pdb.split('/')[-2]
# Record assessment in dat file
with open(dat_file, 'a') as f:
f.write(filename + " " + str(time.time()) + '\n')
# Start assession
pctools.printv('\nAssessing ' + pdb + '...', verbosity)
# Reject files larger than 10Mb
file_size = os.stat(pdb).st_size / 1048576
pctools.printv(
'File size: ' + clrs['c'] + '{0:.1g}'.format(file_size) + ' Mb' +
clrs['n'], verbosity)
if file_size > 2:
pctools.printv(clrs['r'] + "File size too large!" + clrs['n'],
verbosity)
pctools.printv(
clrs['y'] +
"Will try to fetch sequences from asymmetric unit." +
clrs['n'], verbosity)
try:
alternative_pdb = os.path.join(
pdb_archive, subfolder,
'pdb' + filename.split('.')[0] + '.ent.gz')
pdb_code, structure, nchains = pctools.parse_pdb_structure(
alternative_pdb)
structure, chain_correspondences[
pdb_code] = pctools.split_states(structure)
nchainspostsplit, seqs, chain_ids = pctools.extract_seqs(
structure, 0)
# Write in fasta file
pctools.printv(
clrs['y'] + "Recording large-pdb sequence" + clrs['n'],
verbosity)
record_fasta(pdb_code,
seqs,
chain_ids,
subfolder,
type='largepdb')
except:
pctools.printv(
clrs['r'] + "Failed to fetch sequence!" + clrs['n'],
verbosity)
continue
try:
pdb_code, structure, nchains = pctools.parse_pdb_structure(pdb)
pctools.printv(
'Number of chains in structure ' + clrs['y'] + pdb_code +
clrs['n'] + ': ' + str(nchains), verbosity)
# Reject structures with more than 60 chains
if int(nchains) > 60:
pctools.printv(
"Number of chains (" + clrs['y'] + str(nchains) +
clrs['n'] + ") larger than 60! " + clrs['r'] +
"Too many chains!" + clrs['n'], verbosity)
pctools.printv(
clrs['y'] + "Will try to fetch sequences anyway." +
clrs['n'], verbosity)
try:
pdb_code, structure, nchains = pctools.parse_pdb_structure(
pdb)
structure, chain_correspondences[
pdb_code] = pctools.split_states(structure)
nchainspostsplit, seqs, chain_ids = pctools.extract_seqs(
structure, 0)
pctools.printv(
clrs['y'] + "Recording large-pdb sequence" + clrs['n'],
verbosity)
# Write in fasta file
record_fasta(pdb_code,
seqs,
chain_ids,
subfolder,
type='largepdb')
except:
pctools.printv(
clrs['r'] + "Failed to fetch sequence!" + clrs['n'],
verbosity)
continue
structure, chain_correspondences[pdb_code] = pctools.split_states(
structure)
nchainspostsplit, seqs, chain_ids = pctools.extract_seqs(
structure, 0)
pctools.printv(
'Number of chains (' + clrs['c'] + str(nchains) + clrs['n'] +
') and file size (' + clrs['c'] + str(file_size) + clrs['n'] +
') OK.' + clrs['g'] + ' Proceeding.' + clrs['n'] + '\n',
verbosity)
# Try to get info from the canonic pdb header (homonimous to pdb1)
canonpdb = "pdb" + pdb_code + ".ent.gz"
try:
contents = pctools.parse_pdb_contents(
os.path.join(pdb_archive, subfolder, canonpdb))[1]
except:
pctools.printv(
clrs['r'] +
'\n\n Mismatch between pdb and biounit entries...' +
clrs['n'], verbosity)
author, software = pctools.get_annotated_states(contents)
pctools.printv(
'Author determined biological unit = ' + str(author),
verbosity)
pctools.printv(
'Software determined quaternary structure= ' + str(software),
verbosity)
# Start assessing sequences and structures (from 2 up to 26 chains)
if 1 < int(nchains) < 61:
ids, proteinpair = pctools.get_pairwise_ids(seqs, nchains)
for id in ids:
if id[0] >= 90:
color = clrs['g']
else:
color = clrs['r']
pctools.printv(
'Identity between chains ' + clrs['y'] + str(id[1]) +
clrs['n'] + ' and ' + clrs['y'] + str(id[2]) +
clrs['n'] + ' is ' + color + str(id[0]) + "%" +
clrs['n'] + ".", verbosity)
# Save records for pure h**o-oligomers
if all(id[0] > 90 for id in ids) and proteinpair is True:
pctools.printv(
"All identities over 90%. Likely " + clrs['b'] +
"h**o-oligomeric" + clrs['n'] + ".", verbosity)
pctools.printv(clrs['y'] + "FETCHING" + clrs['n'] + ".\n",
verbosity)
# Write file to database
newfile = os.path.join(pdb_homo_archive, subfolder,
pdb_code + ".pdb")
if not os.path.isdir(
os.path.join(pdb_homo_archive, subfolder)):
os.mkdir(os.path.join(pdb_homo_archive, subfolder))
io.set_structure(structure)
io.save(newfile)
pctools.gzip_pdb(newfile)
# Write to log file
with open(log_file, 'a') as f:
f.write(
str(pdb_code) + "," + str(nchains) + "," +
'/'.join(author) + "," + '/'.join(software) + "," +
str(os.path.getctime(newfile + '.gz')) + '\n')
# Write in fasta file
pctools.printv(
clrs['y'] + "Recording h**o-oligomer sequence." +
clrs['n'], verbosity)
record_fasta(pdb_code,
seqs,
chain_ids,
subfolder,
type='h**o')
# Investigate partial h**o-oligomers
elif any(id[0] > 90 for id in ids) and proteinpair is True:
at_least_one_interface = False
for id in ids:
if id[0] > 90:
# Check if similar chains share interfaces
if pctools.check_interfaces(
structure, id[1], id[2]):
at_least_one_interface = True
pctools.printv(
'Contacts found between chains ' +
clrs['g'] + str(id[1]) + clrs['n'] +
' and ' + clrs['g'] + str(id[2]) +
clrs['n'] + ' sharing ' + clrs['g'] +
str(id[0]) + clrs['n'] + " % identity.",
verbosity)
pctools.printv(
"At least one putative " + clrs['b'] +
"h**o-oligomeric " + clrs['n'] +
"interface found.", verbosity)
pctools.printv(
clrs['y'] + "FETCHING" + clrs['n'] + ".\n",
verbosity)
# Write file to database
newfile = os.path.join(pdb_homo_archive,
subfolder,
pdb_code + ".pdb")
if not os.path.isdir(
os.path.join(pdb_homo_archive,
subfolder)):
os.mkdir(
os.path.join(pdb_homo_archive,
subfolder))
io.set_structure(structure)
io.save(newfile)
pctools.gzip_pdb(newfile)
# Write to log file
with open(log_file, 'a') as f:
f.write(
str(pdb_code) + "," + str(nchains) +
"," + '/'.join(author) + "," +
'/'.join(software) + "," +
str(os.path.getctime(newfile +
'.gz')) + '\n')
# Write in fasta file
pctools.printv(
clrs['y'] +
"Recording h**o-oligomer sequence." +
clrs['n'], verbosity)
record_fasta(pdb_code,
seqs,
chain_ids,
subfolder,
type='h**o')
break
if at_least_one_interface is False:
pctools.printv(
"No h**o-oligomeric interface found. Likely " +
clrs['r'] + "hetero-oligomeric" + clrs['n'] + ".",
verbosity)
pctools.printv(
clrs['y'] + "Recording hetero-oligomer sequence" +
clrs['n'], verbosity)
# Write in fasta file
record_fasta(pdb_code,
seqs,
chain_ids,
subfolder,
type='hetero')
elif proteinpair is False:
pctools.printv(
clrs['r'] + "No proteic chain pairs found" +
clrs['n'] + ".", verbosity)
if any([set(seq[1]) != {'X'} for seq in seqs]):
pctools.printv(
clrs['y'] + "Protein sequences found though" +
clrs['n'], verbosity)
pctools.printv(
clrs['y'] + "Recording hetero-oligomer sequence" +
clrs['n'], verbosity)
# Write in fasta file
record_fasta(pdb_code,
seqs,
chain_ids,
subfolder,
type='hetero')
else:
pctools.printv(
clrs['r'] +
"Not even a single protein chain. Disregarding." +
clrs['n'], verbosity)
else:
pctools.printv(
"No similar chains found. Likely " + clrs['r'] +
"hetero-oligomeric" + clrs['n'] + ".", verbosity)
pctools.printv(
clrs['y'] + "Recording hetero-oligomer sequence" +
clrs['n'], verbosity)
record_fasta(pdb_code,
seqs,
chain_ids,
subfolder,
type='hetero')
elif int(nchains) == 1:
pctools.printv(
"Only one chain found. Likely " + clrs['r'] + "monomeric" +
clrs['n'] + ".", verbosity)
pctools.printv(
clrs['y'] + "Recording monomer sequence." + clrs['n'],
verbosity)
structure, chain_correspondences[
pdb_code] = pctools.split_states(structure)
nchains, seqs, chain_ids = pctools.extract_seqs(structure, 0)
record_fasta(pdb_code, seqs, chain_ids, subfolder, type='mono')
except:
errtype, errvalue, errtraceback = sys.exc_info()
errtypeshort = str(errtype).split('\'')[1]
pctools.printv(
clrs['r'] + '*' + str(errtypeshort) + ': ' + str(errvalue) +
' l.' + str(errtraceback.tb_lineno) + '*' + clrs['n'],
verbosity)
traceback.print_exception(*sys.exc_info())
if errtypeshort == 'KeyboardInterrupt':
quit()
#pctools.printv(clrs['r']+"UNKNOWN FAULT"+clrs['n']+".", verbosity)
if not os.path.isfile(err_file):
with open(err_file, 'w+') as f:
pass
with open(err_file, 'a') as f:
f.write(filename + '\n')
continue
with open(chain_correspondences_file, 'wb') as p:
pickle.dump(chain_correspondences, p, protocol=pickle.HIGHEST_PROTOCOL)
if not os.path.isfile(err_file):
with open(err_file, 'w+') as f:
f.write('\nNo errors. Assessment terminated succesfully.\n')
def score_alignment(alignment_file):
print(clrs['b'] + 'SCORING ALIGNMENT' + clrs['n'] + ' in ' + clrs['y'] +
os.path.basename(alignment_file) + clrs['n'] + '\n')
sequences = list(SeqIO.parse(alignment_file, "pir"))
query_chains = str(sequences[0].seq).split('/')
template_chains = str(sequences[1].seq).split('/')
trimmed_query_chains = []
trimmed_template_chains = []
for query_chain, template_chain in zip(query_chains, template_chains):
leading_gaps = 0
for r in query_chain:
if r == '-':
leading_gaps += 1
else:
break
trailing_gaps = 0
for r in query_chain[::-1]:
if r == '-':
trailing_gaps += 1
else:
break
if trailing_gaps == 0:
trimmed_query_chains.append(query_chain[leading_gaps:])
trimmed_template_chains.append(template_chain[leading_gaps:])
else:
trimmed_query_chains.append(
query_chain[leading_gaps:-trailing_gaps])
trimmed_template_chains.append(
template_chain[leading_gaps:-trailing_gaps])
relative_wscores = []
relative_scores = []
for q_chain, t_chain in zip(trimmed_query_chains, trimmed_template_chains):
pctools.printv(
'\nCalculating ' + clrs['y'] + 'maximum scores' + clrs['n'] +
' for chain segments:', g_args.verbosity)
max_score, max_wscores = score_pairwise(t_chain, t_chain,
MatrixInfo.blosum62, 0, 0)
pctools.printv(
'\nCalculating ' + clrs['y'] + 'actual scores' + clrs['n'] +
' for chain segments:', g_args.verbosity)
score, wscores = score_pairwise(q_chain, t_chain, MatrixInfo.blosum62,
0, 0)
relative_scores.append(round(score * 100 / max_score, 2))
for max_wscore, wscore in zip(max_wscores, wscores):
if max_wscore != 0:
relative_wscore = round(wscore * 100 / max_wscore, 2)
else:
relative_wscore = 100
relative_wscores.append(relative_wscore)
relative_score = sum(relative_scores) / len(relative_scores)
string = ''
for relative_wscore in relative_wscores:
if relative_wscore > g_args.similarity_cutoff:
color = 'g'
else:
color = 'r'
if string == '':
string += (clrs[color] + str(relative_wscore) + clrs['n'])
else:
string += (' ~ ' + clrs[color] + str(relative_wscore) + clrs['n'])
print('\nRelative score per 30-res segment: ' + string + clrs['n'])
return relative_score, relative_wscores, len(query_chains)
def generate_ali(alignments, best_oligo_template, residue_index_mapping, args):
best_oligo_template_file = best_oligo_template + "_CHOIR_RenamedChainsTemplate"
final_alignment = os.path.join(
workdir,
input_name + '_' + best_oligo_template + '_CHOIR_Alignment.ali')
getseq = False
alignment_dict = {}
full_residue_mapping = {}
# Parse individual GESAMT alignments and organize in a per-chain dictionary
for fasta_alignment in alignments:
getseq = False
template = False
chain = None
entryseq_dict = {}
for line in open(fasta_alignment, 'r').readlines():
# Only record sequence if line above starts with >
if getseq is True:
getseq = False
seq = line.replace('\n', '')
# If this is the template, count leading and trailing gaps
if template is True:
template = False
leading_gaps = 0
for r in seq:
if r == '-':
leading_gaps += 1
else:
break
trailing_gaps = 0
for r in seq[::-1]:
if r == '-':
trailing_gaps += 1
else:
break
assert seq is not None, 'Sequence is None'
assert seq != '', 'Sequence is empty'
entryseq_dict[entry] = seq.upper()
del seq
# If it is an entry line, get details and expect sequence
if line.startswith('>'):
entry = line.split('>')[1].split('(')[0].split(
'.pdb')[0].replace('\n', '')
# If entry is template, use chain as reference
if entry == best_oligo_template_file:
chain = line.split('(')[1].split(')')[0]
template = True
getseq = True
# Remove leading and trailing gaps from the alignment for both template and query
if trailing_gaps == 0:
for entry, seq in entryseq_dict.items():
entryseq_dict[entry] = leading_gaps * '-' + seq[leading_gaps:]
else:
for entry, seq in entryseq_dict.items():
entryseq_dict[entry] = leading_gaps * '-' + seq[
leading_gaps:-trailing_gaps] + trailing_gaps * '-'
if residue_index_mapping is not None:
full_residue_mapping[chain] = collections.OrderedDict()
for res, i in residue_index_mapping.items():
full_residue_mapping[chain][res] = i + leading_gaps
else:
full_residue_mapping[chain] = leading_gaps
alignment_dict[chain] = entryseq_dict
pctools.printv(
'Removed ' + clrs['c'] + str(leading_gaps) + clrs['n'] +
' leading gaps and ' + clrs['c'] + str(trailing_gaps) + clrs['n'] +
' trailing gaps from chain ' + clrs['c'] + chain + clrs['n'] +
' alignment.\n', verbosity)
# If symmetry is desired, reduce all chains to match the size of the smallest
if args.symmetry:
max_leading_gaps = 0
max_trailing_gaps = 0
for chain, seqs in alignment_dict.items():
for entry, seq in seqs.items():
if entry == best_oligo_template_file:
leading_gaps = 0
for r in seq:
if r == '-':
leading_gaps += 1
else:
break
if leading_gaps > max_leading_gaps:
max_leading_gaps = leading_gaps
trailing_gaps = 0
for r in seq[::-1]:
if r == '-':
trailing_gaps += 1
else:
break
if trailing_gaps > max_trailing_gaps:
max_trailing_gaps = trailing_gaps
pctools.printv(
'To cope with symmetry restraints, the modelled sequence will contain '
+ clrs['c'] + str(max_leading_gaps) + clrs['n'] +
' leading gaps and ' + clrs['c'] + str(max_trailing_gaps) +
clrs['n'] + ' trailing gaps' + clrs['n'] + '.\n', verbosity)
print(max_trailing_gaps)
for chain, seqs in alignment_dict.items():
if max_trailing_gaps == 0:
seqs[entry] = max_leading_gaps * '-' + seqs[entry][
max_leading_gaps:]
else:
seqs[entry] = max_leading_gaps * '-' + seqs[
entry][max_leading_gaps:
-max_trailing_gaps] + max_trailing_gaps * '-'
# Find out first and last chains
first_chain = sorted(alignment_dict)[0]
last_chain = sorted(alignment_dict)[-1]
# Create strings to write in alignment file
alignment_string_dict = {}
for entry in [input_name, best_oligo_template_file]:
if entry == input_name:
alignment_string_dict[
entry] = ">P1;" + input_name + "\nsequence:" + input_name + ":FIRST:" + first_chain + ":LAST:" + last_chain + "::::\n"
elif entry == best_oligo_template_file:
alignment_string_dict[
entry] = ">P1;" + best_oligo_template_file + ".pdb\nstructureX:" + best_oligo_template_file + ".pdb:FIRST:" + first_chain + ":LAST:" + last_chain + "::::\n"
for chain, entryseq in sorted(alignment_dict.items()):
if chain == last_chain:
alignment_string_dict[entry] += entryseq[entry] + '*\n'
else:
alignment_string_dict[entry] += entryseq[entry] + '/\n'
# Write alignment file
with open(final_alignment, 'w') as f:
for entry, entrystring in alignment_string_dict.items():
pctools.printv(entrystring, verbosity)
f.write(entrystring)
print('Modeller Alignment written to ' + clrs['g'] +
os.path.basename(final_alignment) + clrs['n'] + '\n')
return final_alignment, full_residue_mapping