def process_locus(lr, srin, args):
if len(lr) == 0: return None
totalrange = get_total_range(lr)
#print '^^^^ Locus ^^^^'
#print totalrange.get_range_string()
#print str(len(lr))+"\t"+str(len(sr))+"\t"+str(len(srjun))
# Get fuzzys from of all short reads
sr = {}
#do this more time consuming cutdown ont he SR data after sending to a thread
for srgpd in srin:
srfz = GenePredFuzzyBasics.FuzzyGenePred(
srgpd, juntol=args.junction_tolerance)
for j in srfz.fuzzy_junctions:
junstr = j.left.chr + ':' + str(j.left.end) + ',' + str(
j.right.end)
if junstr not in sr:
sr[junstr] = {}
sr[junstr]['cnt'] = 0
sr[junstr]['fzjun'] = j
sr[junstr]['cnt'] += 1
#srfzs = [GenePredFuzzyBasics.FuzzyGenePred(x) for x in srjun]
#for i in range(0,len(srfzs)): srfzs[i].gpds[0].entry['name'] = 'SR_'+str(i)
fzs = GenePredFuzzyBasics.greedy_gpd_list_to_combined_fuzzy_list(
lr, args.junction_tolerance)
#print str(len(fzs)) + " genepreds"
outputs = []
#if args.threads > 1:
# p = Pool(processes=args.threads)
for fz in fzs:
#if args.by_read:
# if args.threads > 1 and args.by_read:
# p.apply_async(do_fuzzy,args=(fz,sr,args),callback=do_outs)
# else:
# outs = do_fuzzy(fz,sr,args)
# do_outs([outs,totalrange])
#else:
outs = do_fuzzy(fz, sr, args)
# do_outs([outs,totalrange])
for o in outs:
outputs.append(o)
#if args.threads > 1 and args.by_read:
# p.close()
# p.join()
#if not args.by_read:
return [outputs, totalrange]
def process_locus(lr, srin, args):
if len(lr) == 0:
return None
totalrange = get_total_range(lr)
# print '^^^^ Locus ^^^^'
# print totalrange.get_range_string()
# print str(len(lr))+"\t"+str(len(sr))+"\t"+str(len(srjun))
# Get fuzzys from of all short reads
sr = {}
# do this more time consuming cutdown ont he SR data after sending to a thread
for srgpd in srin:
srfz = GenePredFuzzyBasics.FuzzyGenePred(srgpd, juntol=args.junction_tolerance)
for j in srfz.fuzzy_junctions:
junstr = j.left.chr + ":" + str(j.left.end) + "," + str(j.right.end)
if junstr not in sr:
sr[junstr] = {}
sr[junstr]["cnt"] = 0
sr[junstr]["fzjun"] = j
sr[junstr]["cnt"] += 1
# srfzs = [GenePredFuzzyBasics.FuzzyGenePred(x) for x in srjun]
# for i in range(0,len(srfzs)): srfzs[i].gpds[0].entry['name'] = 'SR_'+str(i)
fzs = GenePredFuzzyBasics.greedy_gpd_list_to_combined_fuzzy_list(lr, args.junction_tolerance)
# print str(len(fzs)) + " genepreds"
outputs = []
# if args.threads > 1:
# p = Pool(processes=args.threads)
for fz in fzs:
# if args.by_read:
# if args.threads > 1 and args.by_read:
# p.apply_async(do_fuzzy,args=(fz,sr,args),callback=do_outs)
# else:
# outs = do_fuzzy(fz,sr,args)
# do_outs([outs,totalrange])
# else:
outs = do_fuzzy(fz, sr, args)
# do_outs([outs,totalrange])
for o in outs:
outputs.append(o)
# if args.threads > 1 and args.by_read:
# p.close()
# p.join()
# if not args.by_read:
return [outputs, totalrange]
def main():
#do our inputs
args = do_inputs()
global gout
gout = args.output
gls = GenePredBasics.GenePredLocusStream(args.input)
fgs = GenePredFuzzyBasics.FuzzyGenePredSeparator()
if args.threads > 1:
p = Pool(processes=args.threads)
while True:
buffer = gls.read_locus()
if not buffer: break
if args.threads > 1:
p.apply_async(process_buffer,
args=(buffer, args),
callback=out_gpds)
else:
v = process_buffer(buffer, args)
out_gpds(v)
if args.threads > 1:
p.close()
p.join()
sys.stderr.write("\n")
def evaluate_junctions(fz, sr, args):
cnt = 0
source_names = [x.entry['name'] for x in fz.gpds]
working = fz.copy()
if len(working.fuzzy_junctions) == 0: return []
for i in range(0, len(working.fuzzy_junctions)):
newjun = working.fuzzy_junctions[i]
newjun.left.get_payload()['junc'] = []
newjun.right.get_payload()['junc'] = []
oldjun = fz.fuzzy_junctions[i]
for srjun in sr:
sjun = sr[srjun]['fzjun']
if oldjun.overlaps(sjun, args.junction_tolerance):
for i in range(0, min(sr[srjun]['cnt'], args.downsample)):
newjun.left.get_payload()['junc'].append(
sjun.left.get_payload()['junc'][0])
newjun.right.get_payload()['junc'].append(
sjun.right.get_payload()['junc'][0])
cnt += 1
juncs = []
starts = []
ends = []
evidences = []
for i in range(0, len(fz.fuzzy_junctions)):
evidence = len(working.fuzzy_junctions[i].left.get_payload()['junc'])
if evidence >= args.required_evidence:
if i == 0:
starts.append(working.start.start)
elif working.fuzzy_junctions[i].left.get_payload()['start']:
starts.append(working.fuzzy_junctions[i].left.get_payload()
['start'].start)
else:
starts.append(working.fuzzy_junctions[i - 1].right.start)
#now ends
if i == len(fz.fuzzy_junctions) - 1:
ends.append(working.end.end)
elif working.fuzzy_junctions[i].right.get_payload()['end']:
ends.append(
working.fuzzy_junctions[i].right.get_payload()['end'].end)
else:
ends.append(working.fuzzy_junctions[i + 1].left.end)
bestleft = GenePredFuzzyBasics.mode(
working.fuzzy_junctions[i].left.get_payload()['junc'])
bestright = GenePredFuzzyBasics.mode(
working.fuzzy_junctions[i].right.get_payload()['junc'])
juncs.append([bestleft, bestright])
#print 'jun '+str(i)+' evid: '+str(evidence)+" "+str(bestleft)+" "+str(bestright)
else:
starts.append([])
ends.append([])
juncs.append([])
evidences.append(evidence)
#print juncs
#print starts
#print ends
#print evidences
# now we can put together the runs
runs = []
current_run = []
for i in range(0, len(evidences)):
if evidences[i] < args.required_evidence:
if len(current_run) > 0:
runs.append(current_run)
current_run = []
continue
current_run.append(i)
if len(current_run) > 0:
runs.append(current_run)
# now the runs are in runs
#print 'runs:'
parts = []
for run in runs:
sarr = []
sarr.append(starts[run[0]] - 1) #put back to zero index
earr = []
for i in range(0, len(run)):
sarr.append(juncs[run[i]][1] - 1)
earr.append(juncs[run[i]][0])
earr.append(ends[run[-1]])
# ready to build a genepred!
part = ''
part += str(working.start.chr) + "\t"
part += '+' + "\t"
part += str(sarr[0]) + "\t"
part += str(earr[-1]) + "\t"
part += str(sarr[0]) + "\t"
part += str(earr[-1]) + "\t"
part += str(len(sarr)) + "\t"
part += ','.join([str(x) for x in sarr]) + ',' + "\t"
part += ','.join([str(x) for x in earr]) + ','
# Final quality check here
gpd = GenePredEntry("test1\ttest1\t" + part)
if not gpd.is_valid():
sys.stderr.write("\nWARNING skipping invalid GPD\n" +
gpd.get_line() + "\n")
continue
parts.append([part, source_names])
#print parts
return parts
def evaluate_junctions(fz, sr, args):
cnt = 0
source_names = [x.entry["name"] for x in fz.gpds]
working = fz.copy()
if len(working.fuzzy_junctions) == 0:
return []
for i in range(0, len(working.fuzzy_junctions)):
newjun = working.fuzzy_junctions[i]
newjun.left.get_payload()["junc"] = []
newjun.right.get_payload()["junc"] = []
oldjun = fz.fuzzy_junctions[i]
for srjun in sr:
sjun = sr[srjun]["fzjun"]
if oldjun.overlaps(sjun, args.junction_tolerance):
for i in range(0, min(sr[srjun]["cnt"], args.downsample)):
newjun.left.get_payload()["junc"].append(sjun.left.get_payload()["junc"][0])
newjun.right.get_payload()["junc"].append(sjun.right.get_payload()["junc"][0])
cnt += 1
juncs = []
starts = []
ends = []
evidences = []
for i in range(0, len(fz.fuzzy_junctions)):
evidence = len(working.fuzzy_junctions[i].left.get_payload()["junc"])
if evidence >= args.required_evidence:
if i == 0:
starts.append(working.start.start)
elif working.fuzzy_junctions[i].left.get_payload()["start"]:
starts.append(working.fuzzy_junctions[i].left.get_payload()["start"].start)
else:
starts.append(working.fuzzy_junctions[i - 1].right.start)
# now ends
if i == len(fz.fuzzy_junctions) - 1:
ends.append(working.end.end)
elif working.fuzzy_junctions[i].right.get_payload()["end"]:
ends.append(working.fuzzy_junctions[i].right.get_payload()["end"].end)
else:
ends.append(working.fuzzy_junctions[i + 1].left.end)
bestleft = GenePredFuzzyBasics.mode(working.fuzzy_junctions[i].left.get_payload()["junc"])
bestright = GenePredFuzzyBasics.mode(working.fuzzy_junctions[i].right.get_payload()["junc"])
juncs.append([bestleft, bestright])
# print 'jun '+str(i)+' evid: '+str(evidence)+" "+str(bestleft)+" "+str(bestright)
else:
starts.append([])
ends.append([])
juncs.append([])
evidences.append(evidence)
# print juncs
# print starts
# print ends
# print evidences
# now we can put together the runs
runs = []
current_run = []
for i in range(0, len(evidences)):
if evidences[i] < args.required_evidence:
if len(current_run) > 0:
runs.append(current_run)
current_run = []
continue
current_run.append(i)
if len(current_run) > 0:
runs.append(current_run)
# now the runs are in runs
# print 'runs:'
parts = []
for run in runs:
sarr = []
sarr.append(starts[run[0]] - 1) # put back to zero index
earr = []
for i in range(0, len(run)):
sarr.append(juncs[run[i]][1] - 1)
earr.append(juncs[run[i]][0])
earr.append(ends[run[-1]])
# ready to build a genepred!
part = ""
part += str(working.start.chr) + "\t"
part += "+" + "\t"
part += str(sarr[0]) + "\t"
part += str(earr[-1]) + "\t"
part += str(sarr[0]) + "\t"
part += str(earr[-1]) + "\t"
part += str(len(sarr)) + "\t"
part += ",".join([str(x) for x in sarr]) + "," + "\t"
part += ",".join([str(x) for x in earr]) + ","
# Final quality check here
gpd = GenePredEntry("test1\ttest1\t" + part)
if not gpd.is_valid():
sys.stderr.write("\nWARNING skipping invalid GPD\n" + gpd.get_line() + "\n")
continue
parts.append([part, source_names])
# print parts
return parts
def process_buffer(buffer, args):
fzs = GenePredFuzzyBasics.greedy_gpd_list_to_combined_fuzzy_list(
buffer, args.junction_tolerance)
return [fzs, args]
def process_buffer(buffer,args): fzs = GenePredFuzzyBasics.greedy_gpd_list_to_combined_fuzzy_list(buffer,args.junction_tolerance) return [fzs,args]
