def main():
parser = argparse.ArgumentParser(description='Use reference junctions when they are close',formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('--min_intron_size',type=int,default=68,help="INT min intron size")
parser.add_argument('--min_local_support',type=int,default=0,help="INT min number of junctions within search_size of a junction in order to count it")
parser.add_argument('--search_size',type=int,default=10,help="INT search space for reference")
parser.add_argument('--output_fake_psl',help="FASTAFILE reference genome to make a fake PSL output")
parser.add_argument('psl',help="PSLFILENAME or '-' for STDIN")
parser.add_argument('reference_genepred',help="FASTAFILENAME for reference genepred")
args = parser.parse_args()
cpus = multiprocessing.cpu_count()
genome = {}
if args.output_fake_psl:
genome = read_fasta_into_hash(args.output_fake_psl)
#read in the reference genepred first
gpf = GenePredBasics.GenePredFile(args.reference_genepred)
#lets sort entries by chromosome
ref = {}
for e in [x.entry for x in gpf.entries]:
if len(e['exonStarts']) <= 1: continue
if e['chrom'] not in ref:
ref[e['chrom']] = {}
for i in range(1,len(e['exonStarts'])):
if e['exonEnds'][i-1] not in ref[e['chrom']]:
ref[e['chrom']][e['exonEnds'][i-1]] = {}
if e['exonStarts'][i]+1 not in ref[e['chrom']][e['exonEnds'][i-1]]:
ref[e['chrom']][e['exonEnds'][i-1]][e['exonStarts'][i]+1] = e['strand']
#Stored all junctions as 1-base
read_info = {}
pf = GenericFileReader(args.psl)
fcount_total = 0
while True:
line = pf.readline()
if not line: break
if re.match('^#',line): continue
line = line.rstrip()
pe = PSLBasics.line_to_entry(line)
if len(pe['tStarts']) != len(pe['blockSizes']) or len(pe['qStarts']) != len(pe['blockSizes']):
sys.stderr.write("WARNING invalid psl\n")
continue
genepred_line = PSLBasics.convert_entry_to_genepred_line(pe)
ge = GenePredBasics.smooth_gaps(GenePredBasics.line_to_entry(genepred_line),args.min_intron_size)
refjuns = {}
if pe['tName'] in ref: refjuns = ref[pe['tName']]
new_ge = nudge(pe,ge,refjuns,args)
if args.output_fake_psl:
new_psl_line = GenePredBasics.entry_to_fake_psl_line(new_ge,genome)
print new_psl_line
else:
print GenePredBasics.entry_to_line(new_ge)
def main():
parser = argparse.ArgumentParser(formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('genepred',help="FILENAME or use - for STDIN")
parser.add_argument('--smoothing_size',type=int,default=68,help="INT no gaps less than this size")
args = parser.parse_args()
inf = sys.stdin
if args.genepred != '-':
inf = open(args.genepred)
for line in inf:
e = GenePredBasics.line_to_entry(line)
e2 = GenePredBasics.smooth_gaps(e,args.smoothing_size)
print GenePredBasics.entry_to_line(e2)
def parse_gpdfile(tdir,gpdfile,smoothing_factor):
# Go through the long reads and make a genepred
if gpdfile != '-':
fr = FileBasics.GenericFileReader(gpdfile)
else:
fr = sys.stdin
seennames = {}
longreadnumber = 0
of_gpd = open(tdir+'/longreads.gpd','w')
while True:
line = fr.readline()
if not line: break
if re.match('^#',line): #skip comments
continue
longreadnumber += 1
entry = GenePredBasics.smooth_gaps( \
GenePredBasics.line_to_entry(line.rstrip()) \
,smoothing_factor)
readname = entry['name']
if readname in seennames:
sys.stderr.write("Warning: repeat name '"+readname+"'\n")
#set our first name to our bin
entry['name'] = str(longreadnumber)
gline = GenePredBasics.entry_to_line(entry)
of_gpd.write(gline+"\n")
fr.close()
of_gpd.close()
def main():
parser = argparse.ArgumentParser(
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('genepred', help="FILENAME or use - for STDIN")
parser.add_argument('--smoothing_size',
type=int,
default=68,
help="INT no gaps less than this size")
args = parser.parse_args()
inf = sys.stdin
if args.genepred != '-':
inf = open(args.genepred)
for line in inf:
e = GenePredBasics.line_to_entry(line)
e2 = GenePredBasics.smooth_gaps(e, args.smoothing_size)
print GenePredBasics.entry_to_line(e2)
def main():
parser = argparse.ArgumentParser(description="Convert a psl file into a target formated genepred file.")
parser.add_argument('--fill_gaps',type=int,default=0,help="Close gaps this size or smaller.")
parser.add_argument('input_name',help="Input PSL file, use - to indicate STDIN.")
args = parser.parse_args()
pslfilehandle = sys.stdin
if args.input_name != '-':
pslfilehandle = open(args.input_name)
with pslfilehandle as infile:
for line in infile:
psl_entry = PSLBasics.line_to_entry(line)
genepred_line = PSLBasics.convert_entry_to_genepred_line(psl_entry)
if args.fill_gaps > 0:
genepred_entry = GenePredBasics.line_to_entry(genepred_line)
genepred_entry2 = GenePredBasics.smooth_gaps(genepred_entry,args.fill_gaps)
genepred_line = GenePredBasics.entry_to_line(genepred_entry2)
print genepred_line
def check_B_entries(eA,gpdB,overlap,args):
a_unique = True
best_exon_count = 0
best_overlap = 0
best_line = ''
best_frac = 0
ostring = ''
for eB in gpdB.entries:
double_line = GenePredBasics.entry_to_line(eA.entry) + "\t" + GenePredBasics.entry_to_line(eB.entry) + "\n"
gpd_comparison = GenePredBasics.GenePredComparison()
gpd_comparison.set_overlap_requirement(overlap)
if eA.entry['chrom'] != eB.entry['chrom']: continue
# normal is to do full length matches
if not (args.allow_a_subset_of_b_fragments or args.allow_any_fragments):
# do some easy checks
if eA.get_exon_count() != eB.get_exon_count(): continue
gpd_comparison.set_require_all_exons_overlap(True)
gpd_comparison.compare(eA,eB)
if gpd_comparison.output['full_match']:
a_unique = False
if args.output_a_not_in_b:
break # we can bust out of the inner loop if we are only printing stuff unique to a
if not args.best_b_only: # if we aren't waiting for the best, print it
ostring += double_line
else:
# only do the best
if gpd_comparison.output['consecutive_exons'] > best_exon_count \
or (gpd_comparison.output['consecutive_exons'] == best_exon_count \
and gpd_comparison.output['overlap_length'] > best_overlap) \
or (gpd_comparison.output['consecutive_exons'] == best_exon_count \
and gpd_comparison.output['overlap_length'] == best_overlap \
and harmonic_mean(gpd_comparison.output['overlap_fractions']) > best_frac):
best_exon_count = gpd_comparison.output['consecutive_exons']
best_overlap = gpd_comparison.output['overlap_length']
best_line = double_line
best_frac = harmonic_mean(gpd_comparison.output['overlap_fractions'])
# Allow partial matches
else:
gpd_comparison.compare(eA,eB)
if gpd_comparison.output['partial_match']:
# if we require a to be subset of b
if args.allow_a_subset_of_b_fragments \
and not (eA.get_exon_count() < eB.get_exon_count() \
and eA.get_exon_count() == gpd_comparison.output['consecutive_exons']):
break
a_unique = False
if args.output_a_not_in_b:
break
# only do the best
if not args.best_b_only:
ostring += double_line
else:
if gpd_comparison.output['consecutive_exons'] > best_exon_count \
or (gpd_comparison.output['consecutive_exons'] == best_exon_count \
and gpd_comparison.output['overlap_length'] > best_overlap) \
or (gpd_comparison.output['consecutive_exons'] == best_exon_count \
and gpd_comparison.output['overlap_length'] == best_overlap \
and harmonic_mean(gpd_comparison.output['overlap_fractions']) > best_frac):
best_exon_count = gpd_comparison.output['consecutive_exons']
best_overlap = gpd_comparison.output['overlap_length']
best_line = double_line
best_frac = harmonic_mean(gpd_comparison.output['overlap_fractions'])
if best_exon_count > 0 and args.best_b_only:
ostring += best_line
if a_unique and (args.output_a_not_in_b or args.leftouterjoin):
ostring += GenePredBasics.entry_to_line(eA.entry)+"\n"
sys.stdout.write(ostring)
#oval.put(ostring)
return
def main():
parser = argparse.ArgumentParser(
description='Use reference junctions when they are close',
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('--min_intron_size',
type=int,
default=68,
help="INT min intron size")
parser.add_argument(
'--min_local_support',
type=int,
default=0,
help=
"INT min number of junctions within search_size of a junction in order to count it"
)
parser.add_argument('--search_size',
type=int,
default=10,
help="INT search space for reference")
parser.add_argument(
'--output_fake_psl',
help="FASTAFILE reference genome to make a fake PSL output")
parser.add_argument('psl', help="PSLFILENAME or '-' for STDIN")
parser.add_argument('reference_genepred',
help="FASTAFILENAME for reference genepred")
args = parser.parse_args()
cpus = multiprocessing.cpu_count()
genome = {}
if args.output_fake_psl:
genome = read_fasta_into_hash(args.output_fake_psl)
#read in the reference genepred first
gpf = GenePredBasics.GenePredFile(args.reference_genepred)
#lets sort entries by chromosome
ref = {}
for e in [x.entry for x in gpf.entries]:
if len(e['exonStarts']) <= 1: continue
if e['chrom'] not in ref:
ref[e['chrom']] = {}
for i in range(1, len(e['exonStarts'])):
if e['exonEnds'][i - 1] not in ref[e['chrom']]:
ref[e['chrom']][e['exonEnds'][i - 1]] = {}
if e['exonStarts'][i] + 1 not in ref[e['chrom']][e['exonEnds'][i -
1]]:
ref[e['chrom']][e['exonEnds'][i - 1]][e['exonStarts'][i] +
1] = e['strand']
#Stored all junctions as 1-base
read_info = {}
pf = GenericFileReader(args.psl)
fcount_total = 0
while True:
line = pf.readline()
if not line: break
if re.match('^#', line): continue
line = line.rstrip()
pe = PSLBasics.line_to_entry(line)
if len(pe['tStarts']) != len(pe['blockSizes']) or len(
pe['qStarts']) != len(pe['blockSizes']):
sys.stderr.write("WARNING invalid psl\n")
continue
genepred_line = PSLBasics.convert_entry_to_genepred_line(pe)
ge = GenePredBasics.smooth_gaps(
GenePredBasics.line_to_entry(genepred_line), args.min_intron_size)
refjuns = {}
if pe['tName'] in ref: refjuns = ref[pe['tName']]
new_ge = nudge(pe, ge, refjuns, args)
if args.output_fake_psl:
new_psl_line = GenePredBasics.entry_to_fake_psl_line(
new_ge, genome)
print new_psl_line
else:
print GenePredBasics.entry_to_line(new_ge)
def check_B_entries(eA, gpdB, overlap, args):
a_unique = True
best_exon_count = 0
best_overlap = 0
best_line = ''
best_frac = 0
ostring = ''
for eB in gpdB.entries:
double_line = GenePredBasics.entry_to_line(
eA.entry) + "\t" + GenePredBasics.entry_to_line(eB.entry) + "\n"
gpd_comparison = GenePredBasics.GenePredComparison()
gpd_comparison.set_overlap_requirement(overlap)
if eA.entry['chrom'] != eB.entry['chrom']: continue
# normal is to do full length matches
if not (args.allow_a_subset_of_b_fragments
or args.allow_any_fragments):
# do some easy checks
if eA.get_exon_count() != eB.get_exon_count(): continue
gpd_comparison.set_require_all_exons_overlap(True)
gpd_comparison.compare(eA, eB)
if gpd_comparison.output['full_match']:
a_unique = False
if args.output_a_not_in_b:
break # we can bust out of the inner loop if we are only printing stuff unique to a
if not args.best_b_only: # if we aren't waiting for the best, print it
ostring += double_line
else:
# only do the best
if gpd_comparison.output['consecutive_exons'] > best_exon_count \
or (gpd_comparison.output['consecutive_exons'] == best_exon_count \
and gpd_comparison.output['overlap_length'] > best_overlap) \
or (gpd_comparison.output['consecutive_exons'] == best_exon_count \
and gpd_comparison.output['overlap_length'] == best_overlap \
and harmonic_mean(gpd_comparison.output['overlap_fractions']) > best_frac):
best_exon_count = gpd_comparison.output[
'consecutive_exons']
best_overlap = gpd_comparison.output['overlap_length']
best_line = double_line
best_frac = harmonic_mean(
gpd_comparison.output['overlap_fractions'])
# Allow partial matches
else:
gpd_comparison.compare(eA, eB)
if gpd_comparison.output['partial_match']:
# if we require a to be subset of b
if args.allow_a_subset_of_b_fragments \
and not (eA.get_exon_count() < eB.get_exon_count() \
and eA.get_exon_count() == gpd_comparison.output['consecutive_exons']):
break
a_unique = False
if args.output_a_not_in_b:
break
# only do the best
if not args.best_b_only:
ostring += double_line
else:
if gpd_comparison.output['consecutive_exons'] > best_exon_count \
or (gpd_comparison.output['consecutive_exons'] == best_exon_count \
and gpd_comparison.output['overlap_length'] > best_overlap) \
or (gpd_comparison.output['consecutive_exons'] == best_exon_count \
and gpd_comparison.output['overlap_length'] == best_overlap \
and harmonic_mean(gpd_comparison.output['overlap_fractions']) > best_frac):
best_exon_count = gpd_comparison.output[
'consecutive_exons']
best_overlap = gpd_comparison.output['overlap_length']
best_line = double_line
best_frac = harmonic_mean(
gpd_comparison.output['overlap_fractions'])
if best_exon_count > 0 and args.best_b_only:
ostring += best_line
if a_unique and (args.output_a_not_in_b or args.leftouterjoin):
ostring += GenePredBasics.entry_to_line(eA.entry) + "\n"
sys.stdout.write(ostring)
#oval.put(ostring)
return
