def callback(bam, common_count, common_reads, common_cols):
# gather constant reads
const_count = 0
for span in const_spans:
starts = []
ends = []
for start, end in span:
starts.append(start)
ends.append(end)
count, reads = _fetch_reads(bam, gene.chrom, gene.strand if self.stranded else None, starts, ends, self.multiple, False, self.whitelist, self.blacklist, self.uniq_only, self.library_type)
const_count += count
#find counts for each region
for num, start, end, const, names in gene.regions:
count, reads = _fetch_reads(bam, gene.chrom, gene.strand if self.stranded else None, [start], [end], self.multiple, False, self.whitelist, self.blacklist, self.uniq_only, self.library_type)
excl_count, excl_reads = _fetch_reads_excluding(bam, gene.chrom, gene.strand if self.stranded else None, start, end, self.multiple, self.whitelist, self.blacklist, self.library_type)
# remove reads that exclude this region
for read in excl_reads:
if read in reads:
reads.remove(read)
count = count - 1
# find reads that *arent'* in this region
other_reads = 0
for read in common_reads:
if not read in reads and not read in excl_reads:
other_reads += 1
if other_reads > 0:
altindex = float(count - excl_count) / other_reads
else:
altindex = ''
if len(common_reads) > 0:
incl_pct = float(count) / len(common_reads)
excl_pct = float(excl_count) / len(common_reads)
else:
incl_pct = ''
excl_pct = ''
cols = common_cols[:]
cols.append(start)
cols.append(end)
cols.append(const_count)
cols.append(num)
cols.append('const' if const else 'alt')
cols.append(count)
cols.append(excl_count)
cols.append(incl_pct)
cols.append(excl_pct)
cols.append(altindex)
yield cols
def count(self,
bam,
library_type,
coverage=False,
uniq_only=False,
fpkm=False,
norm='',
multiple='complete',
whitelist=None,
blacklist=None,
out=sys.stdout,
quiet=False,
start_only=False):
# This is a separate count implementation because for repeat families,
# we need to combine the counts from multiple regions in the genome,
# so the usual chrom, starts, ends loop breaks down.
stranded = library_type in ['FR', 'RF']
if coverage:
sys.stderr.write(
'Coverage calculations not supported with repeatmasker family models\n'
)
sys.exit(1)
if norm and norm not in ['all', 'mapped']:
sys.stderr.write(
'Normalization "%s" not supported with repeatmasker family models\n'
% norm)
sys.exit(1)
# multireads = set()
# single_count = 0
repeats = {}
total_count = 0.0
for family, member, chrom, start, end, strand in _repeatreader(
self.fname):
if not (family, member) in repeats:
repeats[(family, member)] = {'count': 0, 'size': 0}
if not (family, '*') in repeats:
repeats[(family, '*')] = {'count': 0, 'size': 0}
if not chrom in bam.references:
continue
size = end - start
repeats[(family, '*')]['size'] += size
repeats[(family, member)]['size'] += size
count, reads = _fetch_reads(bam,
chrom,
strand if stranded else None, [start],
[end],
multiple,
False,
whitelist,
blacklist,
library_type=library_type)
repeats[(family, '*')]['count'] += count
repeats[(family, member)]['count'] += count
total_count += count
# for read in reads:
# if read.tags and 'IH' in read.tags:
# ih = int(read.opt('IH'))
# else:
# ih = 1
# if ih == 1:
# single_count += 1
# else:
# multireads.add(read.qname)
sys.stderr.write('Calculating normalization...')
norm_val = None
norm_val_orig = None
if norm == 'all':
norm_val_orig = _find_mapped_count(bam, whitelist, blacklist)
elif norm == 'mapped':
# norm_val_orig = single_count + len(multireads)
norm_val_orig = total_count
if norm_val_orig:
norm_val = float(norm_val_orig) / 1000000
sys.stderr.write('\n')
out.write('## input%s%s\n' %
(' ' if bam.filename else '', bam.filename))
out.write('## model %s %s\n' % (self.get_name(), self.get_source()))
out.write('## library %s\n' % library_type)
out.write('## multiple %s\n' % multiple)
if norm_val:
out.write('## norm %s %s\n' % (norm, float(norm_val_orig)))
out.write('## CPM-factor %s\n' % norm_val)
out.write('#')
out.write('\t'.join(self.get_headers()))
out.write('\tlength\tcount')
if norm_val:
out.write('\tcount (CPM)')
if fpkm:
out.write('\tRPKM')
out.write('\n')
sortedkeys = []
for k in repeats:
sortedkeys.append(k)
sortedkeys.sort()
for class_level in [True, False]:
for k in sortedkeys:
if class_level and k[1] != '*': # Do class level counts first
continue
elif not class_level and k[1] == '*':
continue
cols = [k[0], k[1], repeats[k]['size'], repeats[k]['count']]
if norm_val:
cols.append(repeats[k]['count'] / norm_val)
if fpkm:
cols.append(repeats[k]['count'] /
(repeats[k]['size'] / 1000.0) / norm_val)
out.write('%s\n' % '\t'.join([str(x) for x in cols]))
def callback(bam, common_count, common_reads, common_cols):
# gather constant reads
const_count = 0
for span in const_spans:
starts = []
ends = []
for start, end in span:
starts.append(start)
ends.append(end)
count, reads = _fetch_reads(
bam, gene.chrom,
gene.strand if self.stranded else None, starts, ends,
self.multiple, False, self.whitelist, self.blacklist,
self.uniq_only, self.library_type)
const_count += count
#find counts for each region
for num, start, end, const, names in gene.regions:
count, reads = _fetch_reads(
bam, gene.chrom,
gene.strand if self.stranded else None, [start], [end],
self.multiple, False, self.whitelist, self.blacklist,
self.uniq_only, self.library_type)
excl_count, excl_reads = _fetch_reads_excluding(
bam, gene.chrom,
gene.strand if self.stranded else None, start, end,
self.multiple, self.whitelist, self.blacklist,
self.library_type)
# remove reads that exclude this region
for read in excl_reads:
if read in reads:
reads.remove(read)
count = count - 1
# find reads that *arent'* in this region
other_reads = 0
for read in common_reads:
if not read in reads and not read in excl_reads:
other_reads += 1
if other_reads > 0:
altindex = float(count - excl_count) / other_reads
else:
altindex = ''
if len(common_reads) > 0:
incl_pct = float(count) / len(common_reads)
excl_pct = float(excl_count) / len(common_reads)
else:
incl_pct = ''
excl_pct = ''
cols = common_cols[:]
cols.append(start)
cols.append(end)
cols.append(const_count)
cols.append(num)
cols.append('const' if const else 'alt')
cols.append(count)
cols.append(excl_count)
cols.append(incl_pct)
cols.append(excl_pct)
cols.append(altindex)
yield cols
def count(self, bam, library_type, coverage=False, uniq_only=False, fpkm=False, norm='', multiple='complete', whitelist=None, blacklist=None, out=sys.stdout, quiet=False, start_only=False):
# This is a separate count implementation because for repeat families,
# we need to combine the counts from multiple regions in the genome,
# so the usual chrom, starts, ends loop breaks down.
stranded = library_type in ['FR', 'RF']
if coverage:
sys.stderr.write('Coverage calculations not supported with repeatmasker family models\n')
sys.exit(1)
if norm and norm not in ['all', 'mapped']:
sys.stderr.write('Normalization "%s" not supported with repeatmasker family models\n' % norm)
sys.exit(1)
# multireads = set()
# single_count = 0
repeats = {}
total_count = 0.0
for family, member, chrom, start, end, strand in _repeatreader(self.fname):
if not (family, member) in repeats:
repeats[(family, member)] = {'count': 0, 'size': 0}
if not (family, '*') in repeats:
repeats[(family, '*')] = {'count': 0, 'size': 0}
if not chrom in bam.references:
continue
size = end - start
repeats[(family, '*')]['size'] += size
repeats[(family, member)]['size'] += size
count, reads = _fetch_reads(bam, chrom, strand if stranded else None, [start], [end], multiple, False, whitelist, blacklist, library_type=library_type)
repeats[(family, '*')]['count'] += count
repeats[(family, member)]['count'] += count
total_count += count
# for read in reads:
# if read.tags and 'IH' in read.tags:
# ih = int(read.opt('IH'))
# else:
# ih = 1
# if ih == 1:
# single_count += 1
# else:
# multireads.add(read.qname)
sys.stderr.write('Calculating normalization...')
norm_val = None
norm_val_orig = None
if norm == 'all':
norm_val_orig = _find_mapped_count(bam, whitelist, blacklist)
elif norm == 'mapped':
# norm_val_orig = single_count + len(multireads)
norm_val_orig = total_count
if norm_val_orig:
norm_val = float(norm_val_orig) / 1000000
sys.stderr.write('\n')
out.write('## input%s%s\n' % (' ' if bam.filename else '', bam.filename))
out.write('## model %s %s\n' % (self.get_name(), self.get_source()))
out.write('## library %s\n' % library_type)
out.write('## multiple %s\n' % multiple)
if norm_val:
out.write('## norm %s %s\n' % (norm, float(norm_val_orig)))
out.write('## CPM-factor %s\n' % norm_val)
out.write('#')
out.write('\t'.join(self.get_headers()))
out.write('\tlength\tcount')
if norm_val:
out.write('\tcount (CPM)')
if fpkm:
out.write('\tRPKM')
out.write('\n')
sortedkeys = []
for k in repeats:
sortedkeys.append(k)
sortedkeys.sort()
for class_level in [True, False]:
for k in sortedkeys:
if class_level and k[1] != '*': # Do class level counts first
continue
elif not class_level and k[1] == '*':
continue
cols = [k[0], k[1], repeats[k]['size'], repeats[k]['count']]
if norm_val:
cols.append(repeats[k]['count'] / norm_val)
if fpkm:
cols.append(repeats[k]['count'] / (repeats[k]['size'] / 1000.0) / norm_val)
out.write('%s\n' % '\t'.join([str(x) for x in cols]))