def parse_refgpd(tdir,geneprednames,simplenames):
# get the reference genepreds ready to use in work
column_number = 0
entry_number = 0
of_entries = open(tdir+"/entries.txt",'w')
for file in geneprednames:
column_number += 1
of_ref = open(tdir+"/reference."+str(column_number)+".bed",'w')
gfr = FileBasics.GenericFileReader(file)
while True:
line = gfr.readline()
if not line: break
if re.match('^#',line): continue
entry_number += 1
line = line.rstrip("\n")
entry = GenePredBasics.line_to_entry(line)
entry_length = 0
for i in range(0,len(entry['exonStarts'])): entry_length += entry['exonEnds'][i]-entry['exonStarts'][i]
of_entries.write(str(column_number)+ "\t" + simplenames[column_number-1] + "\t" + str(entry_number) + "\t" + entry['gene_name'] + "\t" + entry['name']+"\t"+str(entry_length)+"\n")
exon_number = 0
for i in range(0,len(entry['exonStarts'])):
exon_number += 1
of_ref.write(entry['chrom'] + "\t" + str(entry['exonStarts'][i]) + "\t" \
+ str(entry['exonEnds'][i]) + "\t" + str(entry_number) + "\t" \
+ entry['gene_name'] + "\t" \
+ entry['name'] + "\t" + str(len(entry['exonStarts'])) + "\t" \
+ entry['strand'] + "\t" + str(exon_number) \
+ "\n")
gfr.close()
of_ref.close()
of_entries.close()
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 get_exons_from_seqs(seqs, d, spcf):
sind = 0
oline = ''
for seq in seqs:
sind += 1
psec = 'P' #primary or secondary
if sind > 1: psec = 'S'
d1 = d.copy()
d1['rname'] = seq[1]
if seq[2] == '+': d1['flag'] = 0
else: d1['flag'] = 16
d1['pos'] = seq[3]
d1['cigar'] = seq[4]
d1['cigar_array'] = SamBasics.parse_cigar(seq[4])
skips = set(['H', 'D', 'N'])
total_length = 0
possible_matches = 0
indels = 0
qstart = 0
if d1['cigar_array'][0]['op'] == 'S':
qstart = d1['cigar_array'][0]['val']
if d1['cigar_array'][0]['op'] == 'H':
qstart = d1['cigar_array'][0]['val']
for ce in d1['cigar_array']:
if ce['op'] not in skips:
total_length += ce['val']
if ce['op'] == 'M': possible_matches += ce['val']
elif ce['op'] == 'I':
indels += ce['val']
elif ce['op'] == 'D' and ce['val'] < 68:
indels += ce['val']
fakeseq = 'N' * total_length
d1['seq'] = fakeseq
nline = SamBasics.entry_to_line(d1)
pline = spcf.convert_line(nline)
pentry = PSLBasics.line_to_entry(pline)
#mismatch_count = -1
#if sind == 1 and args.reference_genome: #for primary alignments we can calculate the number of matches
# for i in range(0,len(pentry['blockSizes'])):
# tseq = spcf.genome[pentry['tName']][pentry['tStarts'][i]:pentry['tStarts'][i]+pentry['blockSizes'][i]]
# qseq = sequence[pentry['qStarts'][i]:pentry['qStarts'][i]+pentry['blockSizes'][i]]
# print pentry['blockSizes'][i]
# print tseq
# print qseq
# for j in range(0,len(tseq)):
# if tseq[j].upper() != qseq[j].upper(): mismatch_count += 1
gline = PSLBasics.convert_entry_to_genepred_line(pentry)
gentry = GenePredBasics.line_to_entry(gline)
gsmooth = GenePredBasics.smooth_gaps(gentry, 68)
for i in range(0, len(gsmooth['exonStarts'])):
oline += gsmooth['chrom'] + "\t" + str(
gsmooth['exonStarts'][i]) + "\t" + str(
gsmooth['exonEnds']
[i]) + "\t" + gsmooth['strand'] + "\t" + gsmooth[
'name'] + "\t" + str(possible_matches) + "\t" + str(
indels) + "\t" + psec + "\t" + str(qstart) + "\n"
return oline
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 get_exons_from_seqs(seqs,d,spcf):
sind = 0
oline = ''
for seq in seqs:
sind+=1
psec = 'P' #primary or secondary
if sind > 1: psec = 'S'
d1 = d.copy()
d1['rname'] = seq[1]
if seq[2] == '+': d1['flag'] = 0
else: d1['flag'] = 16
d1['pos'] = seq[3]
d1['cigar'] = seq[4]
d1['cigar_array'] = SamBasics.parse_cigar(seq[4])
skips = set(['H','D','N'])
total_length = 0
possible_matches = 0
indels = 0
qstart = 0
if d1['cigar_array'][0]['op'] == 'S':
qstart = d1['cigar_array'][0]['val']
if d1['cigar_array'][0]['op'] == 'H':
qstart = d1['cigar_array'][0]['val']
for ce in d1['cigar_array']:
if ce['op'] not in skips:
total_length += ce['val']
if ce['op'] == 'M': possible_matches += ce['val']
elif ce['op'] == 'I':
indels += ce['val']
elif ce['op'] == 'D' and ce['val'] < 68:
indels += ce['val']
fakeseq = 'N'*total_length
d1['seq'] = fakeseq
nline = SamBasics.entry_to_line(d1)
pline = spcf.convert_line(nline)
pentry = PSLBasics.line_to_entry(pline)
#mismatch_count = -1
#if sind == 1 and args.reference_genome: #for primary alignments we can calculate the number of matches
# for i in range(0,len(pentry['blockSizes'])):
# tseq = spcf.genome[pentry['tName']][pentry['tStarts'][i]:pentry['tStarts'][i]+pentry['blockSizes'][i]]
# qseq = sequence[pentry['qStarts'][i]:pentry['qStarts'][i]+pentry['blockSizes'][i]]
# print pentry['blockSizes'][i]
# print tseq
# print qseq
# for j in range(0,len(tseq)):
# if tseq[j].upper() != qseq[j].upper(): mismatch_count += 1
gline = PSLBasics.convert_entry_to_genepred_line(pentry)
gentry = GenePredBasics.line_to_entry(gline)
gsmooth = GenePredBasics.smooth_gaps(gentry,68)
for i in range(0,len(gsmooth['exonStarts'])):
oline += gsmooth['chrom'] + "\t" + str(gsmooth['exonStarts'][i])+"\t"+str(gsmooth['exonEnds'][i])+"\t"+gsmooth['strand']+"\t"+gsmooth['name']+"\t"+str(possible_matches)+"\t"+str(indels)+"\t"+psec+"\t"+str(qstart)+"\n"
return oline
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 read_from_fasta_and_genepred(self,genomefastafile,genepredfile):
# read in our genome
seen_names = {}
seen_coords = {}
genepred = {}
with open(genepredfile) as inf:
for line in inf:
if re.match('^#',line): continue
e = GenePredBasics.line_to_entry(line)
hexcoord = hashlib.sha1(e['chrom']+"\t"+e['strand'] + "\t" + str(e['exonStarts'])+"\t" + str(e['exonEnds'])).hexdigest()
#print hex
#print e['gene_name']
#print e['name']
dupname = 0
dupcoord = 0
if hexcoord in seen_coords:
sys.stderr.write("Warning "+ e['name'] + " " + e['gene_name'] + " exists at identical coordinates as another entry\n")
dupcoord = 1
seen_coords[hexcoord] = 1
currname = e['name']
if e['name'] in seen_names:
if dupcoord == 1:
sys.stderr.write("skipping perfect duplicate of "+e['name']+"\n")
continue
newname = e['name'] + "."+str(len(seen_names[e['name']])+1)
currname = newname
seen_names[e['name']].append(newname)
sys.stderr.write("Warning "+ e['name'] + " " + e['gene_name'] + " is a duplicate name.. renaming to "+newname+ "\n")
dupname = 1
else:
seen_names[e['name']] = []
seen_names[e['name']].append(e['name'])
genepred[currname] = e
#print "reading names and locs"
ref = read_fasta_into_hash(genomefastafile)
#print "converting sequences"
for transcript in genepred:
e = genepred[transcript]
if e['chrom'] in ref:
seq = ''
self.transcript_names[transcript] = genepred[transcript]['name']
for i in range(0,e['exonCount']):
seq += ref[e['chrom']][e['exonStarts'][i]:e['exonEnds'][i]]
if e['strand'] == '-': seq = rc(seq)
self.transcripts[transcript] = seq
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 main():
parser = argparse.ArgumentParser()
parser.add_argument('input_file', help="use - for STDIN")
args = parser.parse_args()
inf = sys.stdin
if args.input_file != '-':
inf = open(args.input_file)
for line in inf:
e = GenePredBasics.line_to_entry(line.rstrip())
matches = 0
qstartslist = []
for i in range(0, len(e['exonStarts'])):
mylen = e['exonEnds'][i] - e['exonStarts'][i]
matches += mylen
qstartslist.append(matches - mylen)
qstarts = ','.join([str(x) for x in qstartslist]) + ','
oline = str(matches) + "\t" # 1
oline += "0\t" # 2
oline += "0\t" # 3
oline += "0\t" # 4
oline += "0\t" # 5
oline += "0\t" # 6
oline += "0\t" # 7
oline += "0\t" # 8
oline += e['strand'] + "\t" # 9
oline += e['name'] + "\t" # 10
oline += str(matches) + "\t" # 11
oline += "0\t" # 12
oline += str(matches) + "\t" # 13
oline += str(e['chrom']) + "\t" # 14
oline += str(e['exonEnds'][-1]) + "\t" # 15
oline += str(e['exonStarts'][0]) + "\t" # 16
oline += str(e['exonEnds'][-1]) + "\t" # 17
oline += str(len(e['exonStarts'])) + "\t" # 18
oline += ','.join([
str(e['exonEnds'][x] - e['exonStarts'][x])
for x in range(0, len(e['exonStarts']))
]) + ',' + "\t" # 19
oline += qstarts + "\t" # 20
oline += ','.join([str(x) for x in e['exonStarts']]) + ',' # 21
print oline
inf.close()
def main():
parser = argparse.ArgumentParser()
parser.add_argument('input_file',help="use - for STDIN")
args = parser.parse_args()
inf = sys.stdin
if args.input_file != '-':
inf = open(args.input_file)
for line in inf:
e = GenePredBasics.line_to_entry(line.rstrip())
matches = 0
qstartslist = []
for i in range(0,len(e['exonStarts'])):
mylen = e['exonEnds'][i]-e['exonStarts'][i]
matches += mylen
qstartslist.append(matches-mylen)
qstarts = ','.join([str(x) for x in qstartslist])+','
oline = str(matches)+"\t" # 1
oline += "0\t" # 2
oline += "0\t" # 3
oline += "0\t" # 4
oline += "0\t" # 5
oline += "0\t" # 6
oline += "0\t" # 7
oline += "0\t" # 8
oline += e['strand']+"\t" # 9
oline += e['name']+"\t" # 10
oline += str(matches)+"\t" # 11
oline += "0\t" # 12
oline += str(matches)+"\t" # 13
oline += str(e['chrom'])+"\t" # 14
oline += str(e['exonEnds'][-1])+"\t" # 15
oline += str(e['exonStarts'][0])+"\t" # 16
oline += str(e['exonEnds'][-1])+"\t" # 17
oline += str(len(e['exonStarts']))+"\t" # 18
oline += ','.join([str(e['exonEnds'][x]-e['exonStarts'][x]) for x in range(0,len(e['exonStarts']))])+','+"\t" # 19
oline += qstarts + "\t" # 20
oline += ','.join([str(x) for x in e['exonStarts']])+',' # 21
print oline
inf.close()
def break_gpdfile(tdir,job_size):
bfcr = BigFileBasics.BigFileChunkReader(tdir+'/longreads.gpd')
bfcr.set_chunk_size_bytes(job_size)
num_jobs = bfcr.chunk_count
for i in range(0,bfcr.chunk_count):
oc = bfcr.open_chunk(i)
job = i+1
of_bed = open(tdir+'/partreads.'+str(job)+'.bed','w')
while True:
line = oc.read_line()
if not line: break
line = line.rstrip("\n")
entry = GenePredBasics.line_to_entry(line)
exon_number = 0
for i in range(0,len(entry['exonStarts'])):
exon_number += 1
of_bed.write(entry['chrom'] + "\t" + str(entry['exonStarts'][i]) + "\t" \
+ str(entry['exonEnds'][i]) + "\t" + entry['name']+"\t" \
+ entry['gene_name'] + "\t" + str(len(entry['exonStarts'])) + "\t" \
+ entry['strand'] + "\t" + str(exon_number) + "\n")
oc.close()
of_bed.close()
return num_jobs
def nudge(psl_entry,gpd_entry,refjun,args):
junctions = []
fcount = 0
if len(gpd_entry['exonStarts']) == 1:
#print "no intron 1"
return gpd_entry
bounds = []
for i in range(1,len(gpd_entry['exonStarts'])):
junc_start = gpd_entry['exonEnds'][i-1]
junc_finish = gpd_entry['exonStarts'][i]+1
bounds.append([junc_start, junc_finish,i-1])
if len(bounds) < 1:
#print "no intron 2"
return gpd_entry
bestbounds = []
for bound in bounds:
best_distance = [10000000,10000000]
best_result = None
for z1 in range(bound[0]-args.search_size,bound[0]+args.search_size+1):
d1 = abs(z1-bound[0])
if z1 in refjun:
for z2 in range(bound[1]-args.search_size,bound[1]+args.search_size+args.search_size+1):
d2 = abs(z2-bound[1])
if z2 in refjun[z1]:
refstrand = refjun[z1][z2]
if d1+d2 < best_distance[0]+best_distance[1]:
best_distance = [d1,d2]
best_result = [z1,z2,refstrand,bound[2]]+best_distance
if best_result:
bestbounds.append(best_result)
if len(bestbounds) < 1:
#nothing fixable
#sys.stderr.write("nothing fixable\n")
return gpd_entry
#Now we have a list of nudgable bounds
#Lets pick a strand
plus_score = 0
minus_score = 0
#print '----'
#print bestbounds
for bound in bestbounds:
if bound[2] == '+':
plus_score += 1/(float(abs(bound[4]))+float(abs(bound[5]))+1)
else:
minus_score += 1/(float(abs(bound[4]))+float(abs(bound[5]))+1)
use_strand = '+'
#print [plus_score,minus_score]
if plus_score < minus_score: use_strand = '-'
#print use_strand
choice_bounds = []
for bound in bestbounds:
if bound[2] == use_strand: choice_bounds.append(bound)
#print '---'
#print GenePredBasics.entry_to_line(gpd_entry)
#print bestbounds
#print choice_bounds
if len(choice_bounds) < 1:
print "ERROR should have choices"
sys.exit()
replacements = {}
for bound in choice_bounds: replacements[bound[3]] = [bound[0],bound[1]]
junctions = []
#print "fixed "+str(len(replacements.keys()))
for i in range(0,len(bounds)):
val = bounds[i]
if i in replacements:
#sys.stderr.write("use replacement\n")
val = replacements[i]
fcount += 1
junctions.append([val[0],val[1]])
#print junctions
#sys.stderr.write("replace\n")
#print junctions
new_gpd_line = gpd_entry['gene_name'] + "\t"
new_gpd_line += gpd_entry['name'] + "\t"
new_gpd_line += gpd_entry['chrom'] + "\t"
new_gpd_line += gpd_entry['strand'] + "\t"
new_gpd_line += str(gpd_entry['txStart']) + "\t"
new_gpd_line += str(gpd_entry['txEnd']) + "\t"
new_gpd_line += str(gpd_entry['cdsStart']) + "\t"
new_gpd_line += str(gpd_entry['cdsEnd']) + "\t"
new_gpd_line += str(len(junctions)+1) + "\t"
exon_starts = [gpd_entry['txStart']]
exon_ends = [] #gpd_entry['txEnd']]
for junc in junctions:
exon_starts.append(junc[1]-1)
exon_ends.append(junc[0])
exon_ends.append(gpd_entry['txEnd'])
new_gpd_line += ','.join([str(x) for x in exon_starts])+','+"\t"
new_gpd_line += ','.join([str(x) for x in exon_ends])+','+"\t"
#print new_gpd_line
new_gpd_entry = GenePredBasics.line_to_entry(new_gpd_line)
#print "got junctions"
#print new_gpd_line
#print '.........'
return new_gpd_entry
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 nudge(psl_entry, gpd_entry, refjun, args):
junctions = []
fcount = 0
if len(gpd_entry['exonStarts']) == 1:
#print "no intron 1"
return gpd_entry
bounds = []
for i in range(1, len(gpd_entry['exonStarts'])):
junc_start = gpd_entry['exonEnds'][i - 1]
junc_finish = gpd_entry['exonStarts'][i] + 1
bounds.append([junc_start, junc_finish, i - 1])
if len(bounds) < 1:
#print "no intron 2"
return gpd_entry
bestbounds = []
for bound in bounds:
best_distance = [10000000, 10000000]
best_result = None
for z1 in range(bound[0] - args.search_size,
bound[0] + args.search_size + 1):
d1 = abs(z1 - bound[0])
if z1 in refjun:
for z2 in range(
bound[1] - args.search_size,
bound[1] + args.search_size + args.search_size + 1):
d2 = abs(z2 - bound[1])
if z2 in refjun[z1]:
refstrand = refjun[z1][z2]
if d1 + d2 < best_distance[0] + best_distance[1]:
best_distance = [d1, d2]
best_result = [z1, z2, refstrand, bound[2]
] + best_distance
if best_result:
bestbounds.append(best_result)
if len(bestbounds) < 1:
#nothing fixable
#sys.stderr.write("nothing fixable\n")
return gpd_entry
#Now we have a list of nudgable bounds
#Lets pick a strand
plus_score = 0
minus_score = 0
#print '----'
#print bestbounds
for bound in bestbounds:
if bound[2] == '+':
plus_score += 1 / (float(abs(bound[4])) + float(abs(bound[5])) + 1)
else:
minus_score += 1 / (float(abs(bound[4])) + float(abs(bound[5])) +
1)
use_strand = '+'
#print [plus_score,minus_score]
if plus_score < minus_score: use_strand = '-'
#print use_strand
choice_bounds = []
for bound in bestbounds:
if bound[2] == use_strand: choice_bounds.append(bound)
#print '---'
#print GenePredBasics.entry_to_line(gpd_entry)
#print bestbounds
#print choice_bounds
if len(choice_bounds) < 1:
print "ERROR should have choices"
sys.exit()
replacements = {}
for bound in choice_bounds:
replacements[bound[3]] = [bound[0], bound[1]]
junctions = []
#print "fixed "+str(len(replacements.keys()))
for i in range(0, len(bounds)):
val = bounds[i]
if i in replacements:
#sys.stderr.write("use replacement\n")
val = replacements[i]
fcount += 1
junctions.append([val[0], val[1]])
#print junctions
#sys.stderr.write("replace\n")
#print junctions
new_gpd_line = gpd_entry['gene_name'] + "\t"
new_gpd_line += gpd_entry['name'] + "\t"
new_gpd_line += gpd_entry['chrom'] + "\t"
new_gpd_line += gpd_entry['strand'] + "\t"
new_gpd_line += str(gpd_entry['txStart']) + "\t"
new_gpd_line += str(gpd_entry['txEnd']) + "\t"
new_gpd_line += str(gpd_entry['cdsStart']) + "\t"
new_gpd_line += str(gpd_entry['cdsEnd']) + "\t"
new_gpd_line += str(len(junctions) + 1) + "\t"
exon_starts = [gpd_entry['txStart']]
exon_ends = [] #gpd_entry['txEnd']]
for junc in junctions:
exon_starts.append(junc[1] - 1)
exon_ends.append(junc[0])
exon_ends.append(gpd_entry['txEnd'])
new_gpd_line += ','.join([str(x) for x in exon_starts]) + ',' + "\t"
new_gpd_line += ','.join([str(x) for x in exon_ends]) + ',' + "\t"
#print new_gpd_line
new_gpd_entry = GenePredBasics.line_to_entry(new_gpd_line)
#print "got junctions"
#print new_gpd_line
#print '.........'
return new_gpd_entry
