return max(groups, key=_auxfun)[0]
print(files)
br = []
rr = []
bc = 0 #count bouke
rc = 0 #count raoul
for x in files:
inter = InterLap()
b = tgt.read_textgrid(mypath + x + "B.TextGrid").tiers[1]
r = tgt.read_textgrid(mypath + x + "R.TextGrid").tiers[1]
bc += len(b)
rc += len(r)
inter.add([convert_to_float(i) for i in r])
tot_overlaps = set()
for i in b:
interval = convert_to_float(i)
overlaps = list(inter.find(interval))
#print(interval[2])
if (len(overlaps) > 0):
overlaps = [tuple(x) for x in overlaps]
for o in overlaps:
tot_overlaps.add(o)
rr.append(most_common([o[2] for o in overlaps]))
br.append(interval[2])
else:
def read_exons(gtf, chrom, cutoff, coverage_array, exclude):
genes = defaultdict(IntervalSet)
splitters = defaultdict(IntervalSet)
interlaps = []
split_iv = InterLap()
# preempt any bugs by checking that we are getting a particular chrom
assert gtf[0] == "|", (
"expecting a tabix query so we can handle chroms correctly")
#f1 = open("selfchaincut.txt","a")
#f2 = open("segdupscut.txt","a")
#f3 = open("coveragecut.txt","a")
for bed in exclude:
# expecting a tabix query so we can handle chroms correctly
a = "|tabix {bed} {chrom}".format(chrom=chrom, bed=bed)
# any file that gets sent in will be used to split regions (just like
# low-coverage). For example, we split on self-chains as well.
#TODO: comment this block if you don't want any filtering by self-chains or segdups
for toks in (
x.strip().split("\t") for x in ts.nopen(a)
): # adds self chains and segdups to splitters list, so that exons can be split, and they are removed from CCRs
s, e = int(toks[1]), int(toks[2])
split_iv.add((s, e))
#if len(toks) > 3:
# f1.write("\t".join(toks)+"\n") # self chain
#else:
# f2.write("\t".join(toks)+"\n") # segdups
for toks in (x.rstrip('\r\n').split("\t") for x in ts.nopen(gtf)
if x[0] != "#"):
if toks[2] not in ("CDS",
"stop_codon") or toks[1] not in ("protein_coding"):
continue
#if toks[0] != "1": break
start, end = map(int, toks[3:5])
gene = toks[8].split('gene_name "')[1].split('"', 1)[0]
assert start <= end, toks
key = toks[0], gene
#cutoff = 0.3
# find sections of exon under certain coverage.
#TODO: comment this if we don't want coverage cutoff filtering
if coverage_array[start - 1:end].min(
) < cutoff: # doesn't bother to run these operations if there is not one bp below the cutoff
#splitters[key].add([(start - 1, end)]) #this takes out the whole exon for one section of poor coverage
a = coverage_array[start - 1:end]
#print str(start-1),end,a
is_under, locs = False, [] # generates "locs" for each exon"
if a[0] < cutoff:
locs.append([start - 1])
is_under = True # so you can initialize is_under
for pos, v in enumerate(
a[1:], start=start
): #enumerates positions in the coverage array starting at the beginning of the exon
if v < cutoff:
if not is_under:
is_under = True
locs.append(
[pos - 1]
) #start, coverage is in bed format, so pos-1 is necessary, since splitters are open left and right side
else:
if is_under:
is_under = False
locs[-1].append(pos) #end
if is_under:
locs[-1].append(
end
) # in this case would end splitter at the end of the exon
splitters[key].add(map(tuple, locs))
#for i in locs:
# f3.write(chrom+"\t"+"\t".join(map(str,i))+"\n")
for s, e in split_iv.find((start - 1, end)):
splitters[key].add([(s, e)])
genes[key].add(
[(start - 1, end)]
) # converts GTF exon coordinates to BED format (subtracts 1 from exon start)
# sort by start so we can do binary search.
genes = dict((k, sorted(v._vals)) for k, v in genes.iteritems())
#ends = dict((k, sorted(v)) for k, v in ends.iteritems())
splits, starts, ends = {}, {}, {}
splitters = dict(splitters)
for chrom_gene, sends in genes.iteritems():
starts[chrom_gene] = [s[0] for s in sends]
ends[chrom_gene] = [s[1] for s in sends]
if chrom_gene in splitters:
splits[chrom_gene] = splitters[chrom_gene]._vals
return starts, ends, splits
