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
