def loadHitRegions(blastFile, minLength, options):
"""
Parse a hit table into a map from read names to lists of (start,end,annot)
"""
hitMap = {}
with InputFile(blastFile) as m8stream:
params = FilterParams.create_from_arguments(options)
hitcount = 0
readcount = 0
keepcount = 0
for (read, hits) in filterM8Stream(m8stream, params,
return_lines=False):
readcount += 1
hitTuples = []
for hit in hits:
hitcount += 1
if abs(hit.qstart - hit.qend) + 1 < minLength:
continue
keepcount += 1
if hit.format == GFF:
annot = "# %d # %d # %s # %s;evalue=%s" % \
(hit.qstart, hit.qend,
hit.strand, hit.hitDesc, hit.evalue)
else:
try:
annot = "%s [%d,%d] %0.1f%% %d bits" % \
(hit.hit, hit.hstart, hit.hend,
hit.pctid, hit.score)
except AttributeError:
annot = "%s [%d,%d] score: %d" % (
hit.hit, hit.hstart, hit.hend, hit.score)
if hit.format == GFF:
reverse = hit.strand != "+"
else:
reverse = hit.hstart > hit.hend
if reverse:
# reverse if hit is backwards
hitTuples.append((hit.qend, hit.qstart, annot))
else:
hitTuples.append((hit.qstart, hit.qend, annot))
hitMap[read] = hitTuples
logging.debug(
"Kept %d of %d hits from %d lines to %d reads" %
(keepcount, hitcount, m8stream.lines, readcount))
return hitMap
def loadHitRegions(blastFile, minLength, options):
"""
Parse a hit table into a map from read names to lists of (start,end,annot)
"""
hitMap = {}
params = FilterParams.create_from_arguments(options)
m8stream = M8Stream(blastFile)
hitcount = 0
readcount = 0
keepcount = 0
for (read, hits) in filterM8Stream(m8stream, params, returnLines=False):
readcount += 1
hitTuples = []
for hit in hits:
hitcount += 1
if abs(hit.qstart - hit.qend) + 1 < minLength:
continue
keepcount += 1
if hit.format == GFF:
annot = "# %d # %d # %s # %s;evalue=%s" % (
hit.qstart, hit.qend, hit.strand, hit.hitDesc, hit.evalue)
else:
try:
annot = "%s [%d,%d] %0.1f%% %d bits" % (
hit.hit, hit.hstart, hit.hend, hit.pctid, hit.score)
except AttributeError:
annot = "%s [%d,%d] score: %d" % (
hit.hit, hit.hstart, hit.hend, hit.score)
if hit.format == GFF:
reverse = hit.strand != "+"
else:
reverse = hit.hstart > hit.hend
if reverse:
# reverse if hit is backwards
hitTuples.append((hit.qend, hit.qstart, annot))
else:
hitTuples.append((hit.qstart, hit.qend, annot))
hitMap[read] = hitTuples
logging.debug(
"Kept %d of %d hits to %d reads" %
(keepcount, hitcount, readcount))
return hitMap
def main():
description = """
Take a blast result table and output a subset of hits based on the
chosen filtering options. If more than one blast file given, use -O
to get multiple output files, otherwise all output data will be
concatenated into one output.
"""
# command line arguments
parser = argparse.ArgumentParser(description=description,
conflict_handler='resolve')
add_hit_table_arguments(parser, flags='all')
parser.add_argument("-o",
"--outfilenome",
dest="outfilename",
default=None,
metavar="OUTFILENAME",
help="Write masked fasta output to OUTFILENAME.")
parser.add_argument(
'-O',
'--autoOutName',
default=False,
action='store_true',
help="Automatically generate output file name from input name "
"and options. Overridden by -o, cannot be used with data "
"from STDIN.")
parser.add_argument('-G',
'--gff',
default=False,
action='store_true',
help="output GFF format instead of input format")
parser.add_argument('hit_table',
nargs='*',
type=argparse.FileType('rU'),
default=[
sys.stdin,
],
help="Table of search results to be filtered. "
"If absent, data will be read from STDIN")
add_universal_arguments(parser)
arguments = parser.parse_args()
setup_logging(arguments)
# check that we have blast file as argument
# if we're not doing auto file names, wriate all outputs to same file
if not arguments.autoOutName:
if arguments.outfilename is not None:
logging.info("Writing data to %s" % (arguments.outfilename))
outfile_handle = open(arguments.outfilename, 'w')
else:
logging.info("writing data to STDOUT")
outfile_handle = sys.stdout
if arguments.gff:
logging.info("Converting to GFF")
# loop over inputs
for infile_handle in arguments.hit_table:
logging.info("reading data from %s" % (infile_handle.name))
if arguments.autoOutName:
outfile_handle = open(getOutputFile(infile_handle.name, arguments),
'w')
# filter
params = FilterParams.create_from_arguments(arguments)
filterM8(infile_handle, outfile_handle, params, to_gff=arguments.gff)
if arguments.autoOutName:
outfile_handle.close()
infile_handle.close()
def main():
description = """
Take a blast result table and output a subset of hits based on the
chosen filtering options. If more than one blast file given, use -O
to get multiple output files, otherwise all output data will be
concatenated into one output.
"""
# command line arguments
parser = argparse.ArgumentParser(
description=description,
conflict_handler='resolve')
add_hit_table_arguments(parser, flags='all')
parser.add_argument(
"-o",
"--outfilenome",
dest="outfilename",
default=None,
metavar="OUTFILENAME",
help="Write masked fasta output to OUTFILENAME.")
parser.add_argument(
'-O',
'--autoOutName',
default=False,
action='store_true',
help="Automatically generate output file name from input name "
"and options. Overridden by -o, cannot be used with data "
"from STDIN.")
parser.add_argument('-G', '--gff', default=False, action='store_true',
help="output GFF format instead of input format")
parser.add_argument('hit_table', nargs='*',
type=argparse.FileType('rU'), default=[sys.stdin, ],
help="Table of search results to be filtered. "
"If absent, data will be read from STDIN")
add_universal_arguments(parser)
arguments = parser.parse_args()
setup_logging(arguments)
# check that we have blast file as argument
# if we're not doing auto file names, wriate all outputs to same file
if not arguments.autoOutName:
if arguments.outfilename is not None:
logging.info("Writing data to %s" % (arguments.outfilename))
outfile_handle = open(arguments.outfilename, 'w')
else:
logging.info("writing data to STDOUT")
outfile_handle = sys.stdout
if arguments.gff:
logging.info("Converting to GFF")
# loop over inputs
for infile_handle in arguments.hit_table:
logging.info("reading data from %s" % (infile_handle.name))
if arguments.autoOutName:
outfile_handle = open(
getOutputFile(
infile_handle.name,
arguments),
'w')
# filter
params = FilterParams.create_from_arguments(arguments)
filterM8(infile_handle, outfile_handle, params, to_gff=arguments.gff)
if arguments.autoOutName:
outfile_handle.close()
infile_handle.close()
def main():
# command line arguments
parser = argparse.ArgumentParser(
description=__doc__,
formatter_class=argparse.RawDescriptionHelpFormatter,
conflict_handler='resolve')
# default to non-overlapping=0
add_hit_table_arguments(parser,
flags='all',
defaults={'nonoverlapping': 0})
parser.add_argument("-o",
"--outfilenome",
dest="outfilename",
default=None,
metavar="OUTFILENAME",
help="Write masked fasta output to OUTFILENAME.")
parser.add_argument('hit_table',
nargs='?',
type=argparse.FileType('rU'),
default=sys.stdin,
help="Table of search results to be filtered. "
"If absent, data will be read from STDIN")
add_universal_arguments(parser)
arguments = parser.parse_args()
setup_logging(arguments)
# output file or STDOUT
if arguments.outfilename is not None:
logging.info("Writing data to %s" % (arguments.outfilename))
outfile_handle = open(arguments.outfilename, 'w')
else:
logging.info("writing data to STDOUT")
outfile_handle = sys.stdout
# input file or STDIN (handled by argparse)
infile_handle = arguments.hit_table
logging.info("reading data from %s" % (infile_handle.name))
# filter, but don't apply nonoverlapping yet
# non-overlapping should be applied per-reference only
params = FilterParams.create_from_arguments(arguments)
# save user supplied value for later
overlap_buffer = params.nonoverlapping
# turn off for now
params.set_nonoverlapping(-1)
# merge
hit_iter = filterM8Stream(infile_handle, params, return_lines=False)
for query, query_hits in hit_iter:
# group by reference hit
hits_by_ref = defaultdict(list)
for hit in query_hits:
hits_by_ref[hit.hit].append(hit)
# one output for query/reference pair
for ref, ref_hits in hits_by_ref.items():
# remove overlaps unless the buffer has been set to <0
if overlap_buffer >= 0:
ref_hits = remove_overlapping_hits(
ref_hits, on_hit=True, buffer=params.nonoverlapping)
ref_hits = remove_overlapping_hits(
ref_hits, on_hit=False, buffer=params.nonoverlapping)
# aggregate values
length, score, identities = 0, 0, 0
for hit in ref_hits:
length += hit.mlen
score += hit.score
try:
# this will be off by 100x
identities += hit.pctid * hit.mlen
except:
# just report pctid=0 if no pctid column in input
pass
outfile_handle.write(
"%s\t%s\t%d\t%d\t%0.2f\n" %
(query, ref, length, score, identities / length))
outfile_handle.close()
infile_handle.close()
