def get_best_homologues(model, chain_ids=None):
"""
Do local BLAST search for homologues, pick the best resolution, fetch models
chain_ids - list of chains to search. If none - do for all of them.
As usual, multi-model is not supported.
Returns dictionary of chain_id: model
"""
# Load data
pdb_info = iotbx.bioinformatics.pdb_info.pdb_info_local()
h = model.get_hierarchy()
result = {}
for chain in h.only_model().chains():
print "Working with chain '%s'" % chain.id
if not chain.is_protein():
print " skipping, not protein. Maybe water or NA"
continue
if chain_ids is not None and chain.id not in chain_ids():
continue
sequence = chain.as_padded_sequence()
l_blast = local_blast.pdbaa(seq=sequence) # why sequence in init???
blast_xml_result = l_blast.run()
# Probably alignment info is lost in this call,
# but we have tools to do alignment, see mmtbx/alignment/__init__.py
blast_summary = summarize_blast_output("\n".join(blast_xml_result))
pdb_ids_to_study = {}
for hit in blast_summary:
# print dir(hit)
hit.show(out=sys.stdout)
# Don't have clear idea what these values mean right now,
# but it would be reasonable to filter somehow.
if hit.identity < 70:
continue
pdb_ids_to_study[hit.pdb_id] = hit.chain_id # can add more info
#
info_list = pdb_info.get_info_list(pdb_ids_to_study.keys())
# It would be good to merge info_list with hits in blast_summary in
# one data structure here and do better filtering.
# Let's go without it for now.
# Sort by resolution in place
info_list.sort(key=lambda tup: tup[1])
best_pdb_id = info_list[0][0]
best_pdb_chain = pdb_ids_to_study[info_list[0][0]]
print "Best pdb:", info_list[0], "chain:", pdb_ids_to_study[
info_list[0][0]]
# print info_list
# Get actual selected model from PDB.
data = fetch(id=best_pdb_id)
m = mmtbx.model.manager(model_input=iotbx.pdb.input(
source_info=None, lines=data.readlines()))
sel = m.selection("chain '%s'" % best_pdb_chain)
result[chain.id] = m.select(sel)
return result
def perfect_pair(sequence, params, pdb_id=None):
result = []
pdb_info = iotbx.bioinformatics.pdb_info.pdb_info_local()
l_blast = local_blast.pdbaa(seq=sequence)
blast_xml_result = l_blast.run(debug=False, binary=params.engine)
try:
blast_summary = summarize_blast_output("\n".join(blast_xml_result))
except StopIteration:
return result
except Exception as e:
if (("mismatched tag" in str(e)) and params.engine == "blastall"):
raise Sorry("setting engine=blastp and try again")
pdb_ids_to_study = {}
for hit in blast_summary:
#hit.show(out=sys.stdout)
hsp = hit.hsp
### The Gapped BLAST algorithm allows gaps (deletions and insertions) to
### be introduced into the alignments
# blastall: the surprising default value (None, None) instead of an integer.
if hsp.gaps == (None, None): hsp.gaps = 0
identity = (hsp.identities - hsp.gaps) * 100 / (hsp.align_length -
hsp.gaps)
if (not params.piece_matching):
ali_identity = len(hsp.query.replace('X', '')) / len(
sequence.replace('X', ''))
identity = identity * ali_identity
if identity < params.identity:
continue
for i in hit.all_ids:
if i[0] == pdb_id: continue
if (i[0] not in pdb_ids_to_study):
pdb_ids_to_study[str(i[0])] = (str(i[1]), identity)
info_lists = pdb_info.get_info_list(pdb_ids_to_study.keys())
info_lists.sort(key=lambda tup: tup[1])
if info_lists:
for info_list in info_lists:
if not info_list[1]: continue
best_pdb_id = info_list[0]
best_pdb_chain = pdb_ids_to_study[info_list[0]][0]
identity = pdb_ids_to_study[info_list[0]][1]
resolution = float(info_list[1])
g_a = group_args(pdb_code=best_pdb_id,
chain_id=best_pdb_chain,
resolution=resolution,
identity=identity)
if params.high_res is None:
result.append(g_a)
result.sort(key=lambda tup: (tup.resolution, -tup.identity))
elif resolution <= params.high_res:
result.append(g_a)
result.sort(key=lambda tup: (-tup.identity, tup.resolution))
result = result[:params.n]
return result
def run(args=(), params=None, out=None):
if (out is None):
out = sys.stdout
if (params is None):
import iotbx.phil
cmdline = iotbx.phil.process_command_line_with_files(
args=args,
master_phil=master_phil,
seq_file_def="blast_pdb.file_name")
params = cmdline.work.extract()
validate_params(params)
params = params.blast_pdb
from iotbx.bioinformatics.structure import get_ncbi_pdb_blast, \
summarize_blast_output
from iotbx.file_reader import any_file
seq_file = any_file(params.file_name,
force_type="seq",
raise_sorry_if_not_expected_format=True,
raise_sorry_if_errors=True)
seq_file.check_file_type("seq")
seq_objects = seq_file.file_object
if (len(seq_objects) == 0):
raise Sorry("Empty sequence file!")
elif (len(seq_objects) > 1):
print("WARNING: multiple sequences provided; searching only the 1st",
file=out)
sequence = seq_objects[0].sequence
if (len(sequence) == 0):
raise Sorry("No data in sequence file.")
elif (len(sequence) < 6):
raise Sorry("Sequence must be at least six residues.")
if (params.output_file is None):
params.output_file = "blast.xml"
blast_out = get_ncbi_pdb_blast(sequence,
file_name=params.output_file,
blast_type=params.blast_type,
expect=params.expect)
print("Wrote results to %s" % params.output_file, file=out)
results = summarize_blast_output(blast_out)
if (len(args) != 0): # command-line mode
print("", file=out)
print("%d matching structures" % len(results), file=out)
print("", file=out)
print("ID Chain evalue length %ident %pos #structures",
file=out)
print("-" * 59, file=out)
for result in results:
result.show(out)
if (len(results) > 0):
return sequence, os.path.abspath(params.output_file)
else:
return sequence, None
def run (args=(), params=None, out=None) :
if (out is None) :
out = sys.stdout
if (params is None) :
import iotbx.phil
cmdline = iotbx.phil.process_command_line_with_files(
args=args,
master_phil=master_phil,
seq_file_def="blast_pdb.file_name")
params = cmdline.work.extract()
validate_params(params)
params = params.blast_pdb
from iotbx.bioinformatics.structure import get_ncbi_pdb_blast, \
summarize_blast_output
from iotbx.file_reader import any_file
seq_file = any_file(params.file_name,
force_type="seq",
raise_sorry_if_not_expected_format=True,
raise_sorry_if_errors=True)
seq_file.check_file_type("seq")
seq_objects = seq_file.file_object
if (len(seq_objects) == 0) :
raise Sorry("Empty sequence file!")
elif (len(seq_objects) > 1) :
print >> out, "WARNING: multiple sequences provided; searching only the 1st"
sequence = seq_objects[0].sequence
if (len(sequence) == 0) :
raise Sorry("No data in sequence file.")
elif (len(sequence) < 6) :
raise Sorry("Sequence must be at least six residues.")
if (params.output_file is None) :
params.output_file = "blast.xml"
blast_out = get_ncbi_pdb_blast(sequence,
file_name=params.output_file,
blast_type=params.blast_type,
expect=params.expect)
print >> out, "Wrote results to %s" % params.output_file
results = summarize_blast_output(blast_out)
if (len(args) != 0) : # command-line mode
print >> out, ""
print >> out, "%d matching structures" % len(results)
print >> out, ""
print >> out, "ID Chain evalue length %ident %pos #structures"
print >> out, "-" * 59
for result in results :
result.show(out)
if (len(results) > 0) :
return sequence, os.path.abspath(params.output_file)
else :
return sequence, None
self.DeleteAllItems()
for result in results:
i = self.InsertStringItem(sys.maxint, result.pdb_id)
self.SetStringItem(i, 1, result.chain_id)
self.SetStringItem(i, 2, "%g" % result.evalue)
self.SetStringItem(i, 3, "%d" % result.length)
self.SetStringItem(i, 4, "%.2f" % result.identity)
self.SetStringItem(i, 5, "%.2f" % result.positives)
def GetSelectedID(self):
item = self.GetFirstSelected()
if (item >= 0):
return self.GetItem(item, 0).GetText()
return None
if (__name__ == "__main__"):
from iotbx.file_reader import any_file
from iotbx.bioinformatics.structure import summarize_blast_output
seq_file = any_file(sys.argv[1], force_type="seq")
seq_file.check_file_type("seq")
seq_objects = seq_file.file_object
assert (len(seq_objects) == 1)
sequence = seq_objects[0].sequence
results = summarize_blast_output(blast_file=sys.argv[2])
app = wx.App(0)
frame = BlastFrame(None, -1, "BLAST results")
frame.SetResults(sequence, results)
frame.Show()
app.MainLoop()
def SetResults (self, results) :
self.DeleteAllItems()
for result in results :
i = self.InsertStringItem(sys.maxint, result.pdb_id)
self.SetStringItem(i, 1, result.chain_id)
self.SetStringItem(i, 2, "%g" % result.evalue)
self.SetStringItem(i, 3, "%d" % result.length)
self.SetStringItem(i, 4, "%.2f" % result.identity)
self.SetStringItem(i, 5, "%.2f" % result.positives)
def GetSelectedID (self) :
item = self.GetFirstSelected()
if (item >= 0) :
return self.GetItem(item, 0).GetText()
return None
if (__name__ == "__main__") :
from iotbx.file_reader import any_file
from iotbx.bioinformatics.structure import summarize_blast_output
seq_file = any_file(sys.argv[1], force_type="seq")
seq_file.check_file_type("seq")
seq_objects = seq_file.file_object
assert (len(seq_objects) == 1)
sequence = seq_objects[0].sequence
results = summarize_blast_output(blast_file=sys.argv[2])
app = wx.App(0)
frame = BlastFrame(None, -1, "BLAST results")
frame.SetResults(sequence, results)
frame.Show()
app.MainLoop()
