class GTFModel(Model):
def __init__(self, fname):
self.fname = fname
Model.__init__(self)
# test for what attributes we need to return
self.gtf = GTF(self.fname)
gene_gen = self.gtf.genes
gene = gene_gen.next()
self.has_isoform = 'isoform_id' in gene.attributes
self.has_biotype = 'gene_biotype' in gene.attributes
def get_source(self):
return self.fname
def get_name(self):
return 'gtf'
def get_headers(self):
out = [
'gene_id',
'gene_name',
]
if self.has_isoform:
out.append('isoform_id')
if self.has_biotype:
out.append('gene_biotype')
out.extend('chrom strand txstart txend'.split())
return out
def get_regions(self):
eta = ETA(self.gtf.fsize(), fileobj=self.gtf)
for gene in self.gtf.genes:
eta.print_status(extra=gene.gene_name)
starts = []
ends = []
# just include all regions - don't worry about transcripts and exons
# the regions encompass all exons anyway...
for num, start, end, const, names in gene.regions:
starts.append(start)
ends.append(end)
out = [
gene.gene_id,
gene.gene_name,
]
if self.has_isoform:
out.append(gene.attributes['isoform_id'] if 'isoform_id' in
gene.attributes else '')
if self.has_biotype:
out.append(gene.attributes['gene_biotype'] if 'gene_biotype' in
gene.attributes else '')
out.extend([gene.chrom, gene.strand, gene.start, gene.end])
yield (gene.chrom, starts, ends, gene.strand, out, None)
eta.done()
class GTFModel(Model):
def __init__(self, fname):
self.fname = fname
Model.__init__(self)
# test for what attributes we need to return
self.gtf = GTF(self.fname)
gene_gen = self.gtf.genes
gene = gene_gen.next()
self.has_isoform = 'isoform_id' in gene.attributes
self.has_biotype = 'gene_biotype' in gene.attributes
def get_source(self):
return self.fname
def get_name(self):
return 'gtf'
def get_headers(self):
out = ['gene_id', 'gene_name', ]
if self.has_isoform:
out.append('isoform_id')
if self.has_biotype:
out.append('gene_biotype')
out.extend('chrom strand txstart txend'.split())
return out
def get_regions(self):
eta = ETA(self.gtf.fsize(), fileobj=self.gtf)
for gene in self.gtf.genes:
eta.print_status(extra=gene.gene_name)
starts = []
ends = []
# just include all regions - don't worry about transcripts and exons
# the regions encompass all exons anyway...
for num, start, end, const, names in gene.regions:
starts.append(start)
ends.append(end)
out = [gene.gene_id, gene.gene_name, ]
if self.has_isoform:
out.append(gene.attributes['isoform_id'] if 'isoform_id' in gene.attributes else '')
if self.has_biotype:
out.append(gene.attributes['gene_biotype'] if 'gene_biotype' in gene.attributes else '')
out.extend([gene.chrom, gene.strand, gene.start, gene.end])
yield (gene.chrom, starts, ends, gene.strand, out, None)
eta.done()
class ExonModel(Model):
def __init__(self, fname):
self.fname = fname
Model.__init__(self)
self.gtf = GTF(self.fname)
gene_gen = self.gtf.genes
gene = gene_gen.next()
self.has_isoform = 'isoform_id' in gene.attributes
self.has_biotype = 'gene_biotype' in gene.attributes
def get_source(self):
return self.fname
def get_name(self):
return 'exon'
def get_headers(self):
out = [
'gene_id',
'gene_name',
]
if self.has_isoform:
out.append('isoform_id')
if self.has_biotype:
out.append('gene_biotype')
out.extend('chrom strand txstart txend'.split())
return out
def get_postheaders(self):
return 'regionstart regionend const_count region_num const_alt count excl_count incl_pct excl_pct alt-index'.split(
)
def get_regions(self):
eta = ETA(self.gtf.fsize(), fileobj=self.gtf)
for gene in self.gtf.genes:
eta.print_status(extra=gene.gene_name)
starts = []
ends = []
const_spans = []
geneout = [
gene.gene_id,
gene.gene_name,
]
if self.has_isoform:
geneout.append(gene.attributes['isoform_id'] if 'isoform_id' in
gene.attributes else '')
if self.has_biotype:
geneout.append(gene.attributes['gene_biotype']
if 'gene_biotype' in gene.attributes else '')
geneout.extend([gene.chrom, gene.strand, gene.start, gene.end])
was_last_const = False
for num, start, end, const, names in gene.regions:
starts.append(start)
ends.append(end)
# assemble a list of lists with contiguous spans of constant regions
# this will let us count junction-spanning reads that are cover two
# constant exons/regions
if const:
if not was_last_const:
const_spans.append([])
const_spans[-1].append((start, end))
was_last_const = True
else:
was_last_const = False
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
yield (gene.chrom, starts, ends, gene.strand, geneout, callback)
eta.done()
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):
self.uniq_only = uniq_only
self.multiple = multiple
self.whitelist = whitelist
self.blacklist = blacklist
self.library_type = library_type
self.stranded = library_type in ['FR', 'RF']
Model.count(self, bam, library_type, coverage, uniq_only, fpkm, norm,
multiple, whitelist, blacklist, out, quiet, start_only)
class ExonModel(Model):
def __init__(self, fname):
self.fname = fname
Model.__init__(self)
self.gtf = GTF(self.fname)
gene_gen = self.gtf.genes
gene = gene_gen.next()
self.has_isoform = 'isoform_id' in gene.attributes
self.has_biotype = 'gene_biotype' in gene.attributes
def get_source(self):
return self.fname
def get_name(self):
return 'exon'
def get_headers(self):
out = ['gene_id', 'gene_name', ]
if self.has_isoform:
out.append('isoform_id')
if self.has_biotype:
out.append('gene_biotype')
out.extend('chrom strand txstart txend'.split())
return out
def get_postheaders(self):
return 'regionstart regionend const_count region_num const_alt count excl_count incl_pct excl_pct alt-index'.split()
def get_regions(self):
eta = ETA(self.gtf.fsize(), fileobj=self.gtf)
for gene in self.gtf.genes:
eta.print_status(extra=gene.gene_name)
starts = []
ends = []
const_spans = []
geneout = [gene.gene_id, gene.gene_name, ]
if self.has_isoform:
geneout.append(gene.attributes['isoform_id'] if 'isoform_id' in gene.attributes else '')
if self.has_biotype:
geneout.append(gene.attributes['gene_biotype'] if 'gene_biotype' in gene.attributes else '')
geneout.extend([gene.chrom, gene.strand, gene.start, gene.end])
was_last_const = False
for num, start, end, const, names in gene.regions:
starts.append(start)
ends.append(end)
# assemble a list of lists with contiguous spans of constant regions
# this will let us count junction-spanning reads that are cover two
# constant exons/regions
if const:
if not was_last_const:
const_spans.append([])
const_spans[-1].append((start, end))
was_last_const = True
else:
was_last_const = False
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
yield (gene.chrom, starts, ends, gene.strand, geneout, callback)
eta.done()
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):
self.uniq_only = uniq_only
self.multiple = multiple
self.whitelist = whitelist
self.blacklist = blacklist
self.library_type = library_type
self.stranded = library_type in ['FR', 'RF']
Model.count(self, bam, library_type, coverage, uniq_only, fpkm, norm, multiple, whitelist, blacklist, out, quiet, start_only)
def get_regions(self):
gtf = GTF(self.fname)
eta = ETA(gtf.fsize(), fileobj=gtf)
for gene in gtf.genes:
eta.print_status(extra=gene.gene_name)
starts = []
ends = []
const_spans = []
was_last_const = False
for num, start, end, const, names in gene.regions:
starts.append(start)
ends.append(end)
# assemble a list of lists with contiguous spans of constant regions
# this will let us count junction-spanning reads that are cover two
# constant exons/regions
if const:
if not was_last_const:
const_spans.append([])
const_spans[-1].append((start, end))
was_last_const = True
else:
was_last_const = False
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
yield (gene.chrom, starts, ends, gene.strand, [gene.gene_name, gene.gene_id, gene.isoform_id, gene.chrom, gene.strand, gene.start, gene.end], callback)
eta.done()
