def plot(self):
# Data #
rows = flatter([(bg.doc + '\n and they assembled', bg.doc + '\n and they are unassembled') for bg in self.parent])
columns = [pg.__doc__ for pg in self.parent.good_barcodes.assembled.children]
data = flatter([([len(pg) for pg in bg.assembled],[len(pg) for pg in bg.unassembled]) for bg in self.parent])
self.frame = pandas.DataFrame(data, index=rows, columns=columns)
# Plot #
fig = pyplot.figure()
colors = ['g','r','y','orange','k']
axes = self.frame.plot(kind='barh', stacked=True, color=colors)
fig = pyplot.gcf()
# Other #
axes.set_title('Primer presence check results for pool %i (0 base pairs mismatches are allowed)' % self.parent.num)
axes.set_xlabel('Number of paired reads')
axes.xaxis.grid(True)
# Save it #
self.save_plot(fig, axes, left=0.15, sep=('x'))
self.frame.to_csv(self.csv_path)
pyplot.close(fig)
def script(self):
"""The script to be submitted to the SLURM queue."""
self.shebang_header = self.shebang_headers[self.language]
self.slurm_header = [self.slurm_headers[k]['tag'] % v for k,v in self.slurm_params.items()]
self.script_header = self.script_headers[self.language]
self.script_footer = self.script_footers[self.language]
return '\n'.join(flatter([self.shebang_header,
self.slurm_header,
self.script_header,
self.command,
self.script_footer]))
def plot(self):
# Data #
rows = ['Pool "%s" (%s)' % (p.long_name, p) for p in self.parent.pools]
columns = flatter([(o.short_name + "_ass", o.short_name + "_unass") for o in self.parent.first.outcomes])
data = [flatter([(len(o.assembled), len(o.unassembled)) for o in pool.outcomes]) for pool in self.parent.pools]
self.frame = pandas.DataFrame(data, index=rows, columns=columns)
# Plot #
fig = pyplot.figure()
colors = ['g','g','gray','gray','y','y','orange','orange','r','r']
axes = self.frame.plot(kind='barh', stacked=True, color=colors)
fig = pyplot.gcf()
# Add pattern #
unass_patches = [p for i,p in enumerate(axes.patches) if (i/len(self.parent))%2 == 1]
for p in unass_patches: p.set_hatch('//')
# Other #
axes.set_title('Assembling reads using PANDAseq')
axes.set_xlabel('Number of paired reads')
axes.xaxis.grid(True)
# Save it #
self.save_plot(fig, axes, sep=('x'))
self.frame.to_csv(self.csv_path)
pyplot.close(fig)
def plot(self):
# Data #
rows = flatter([(bg.doc + '\n and they assembled', bg.doc + '\n and they are unassembled') for bg in self.parent.outcomes])
columns = [pg.__doc__ for pg in self.parent.good_barcodes.assembled.children]
percentage = lambda x,y: 100-(len(x)/len(y))*100 if len(y) != 0 else 0
data_ass = [[percentage(pg.n_filtered, pg.orig_reads) for pg in bg.assembled] for bg in self.parent.outcomes]
data_unass = [[percentage(pg.n_filtered, pg.orig_reads) for pg in bg.unassembled] for bg in self.parent.outcomes]
data = flatter(zip(data_ass,data_unass))
self.frame = pandas.DataFrame(data, index=rows, columns=columns)
# Plot #
fig = pyplot.figure()
colors = ['g','r','y','orange','k']
axes = self.frame.plot(kind='barh', stacked=True, color=colors)
fig = pyplot.gcf()
# Other #
axes.set_title('Fraction of reads discarded because of undetermined base pairs for pool %i' % self.parent.num)
axes.set_xlabel('Percentage of paired reads with "N" (stacked)')
axes.xaxis.grid(True)
# Save it #
self.save_plot(fig, axes, left=0.15, sep=('x'))
self.frame.to_csv(self.csv_path)
pyplot.close(fig)
def script(self):
"""The script to be submitted to the SLURM queue."""
self.shebang_header = self.shebang_headers[self.language]
self.slurm_header = [
self.slurm_headers[k]['tag'] % v
for k, v in self.slurm_params.items()
]
self.script_header = self.script_headers[self.language]
self.script_footer = self.script_footers[self.language]
return '\n'.join(
flatter([
self.shebang_header, self.slurm_header, self.script_header,
self.command, self.script_footer
]))
def plot(self):
# Data #
fwd_pos, rev_pos = self.parent.unassembled.primer_positions
if not fwd_pos and not rev_pos: return
fwd_data = flatter([[k]*v for k,v in fwd_pos.items()])
rev_data = flatter([[k]*v for k,v in rev_pos.items()])
# Plot #
fig = pyplot.figure()
bins = range(min(rev_data + [0]), max(fwd_data + [0])+1, 1)
if fwd_pos: pyplot.hist(fwd_data, bins=bins, histtype='stepfilled', color='b', alpha=0.5, label='Forward')
if rev_pos: pyplot.hist(rev_data, bins=bins, histtype='stepfilled', color='r', alpha=0.5, label='Reverse')
title = "Distribution of primer positions within unassembled sequences"
title += ' (pool %i, group "%s")' % (self.parent.pool.num, self.parent.short_name)
axes = pyplot.gca()
axes.set_title(title)
axes.set_xlabel('Relative position at which the primer is found')
axes.set_ylabel('Number of primers found at this position')
axes.xaxis.grid(False)
axes.legend()
# Save it #
self.save_plot(fig, axes, sep=('y'))
self.json_path = self.path.replace_extension('json')
json.dump({'fwd':fwd_pos, 'rev':rev_pos}, open(self.json_path, 'w'))
pyplot.close(fig)
def stats(self):
result = {}
result['raw'] = tail(self.p.out)
# Special cases #
if "by concatenation not pandaseq" in result['raw']:
result['noalign'] = 0
result['lowqual'] = 0
result['loss'] = 0
return result
# Normal case #
if "pandaseq: error" in result['raw']: raise Exception("Pandaseq did not run properly")
if result['raw'].startswith("ERR\t"): raise Exception("Pandaseq did not run properly")
result['distrib'] = re.findall('STAT\tOVERLAPS\t(.+)$', result['raw'], re.M)
result['distrib'] = map(int, result['distrib'][0].split())
result['lengths'] = flatter([[i+1]*v for i,v in enumerate(result['distrib'])])
result['noalign'] = int(re.findall('STAT\tNOALGN\t(.+)$', result['raw'], re.M)[0])
result['lowqual'] = int(re.findall('STAT\tLOWQ\t(.+)$', result['raw'], re.M)[0])
result['loss'] = 100 * sum(result['distrib'][100:]) / sum(result['distrib'])
return result
def run(self):
# Check samples #
for s in self.samples:
if not s.loaded: s.load()
assert s.clean.exists
assert s.singletons
# Ray needs a non-existing directory otherwise it is unhappy #
self.out_dir = self.p.output_dir
self.out_dir.remove()
# Make the pairs of fastq #
self.paths = lambda s: ('-p', s.clean.fwd.path, s.clean.rev.path, '-s', s.singletons.path)
self.paths = flatter([self.paths(s) for s in self.samples])
# Call Ray on different setting #
if os.environ.get('CSCSERVICE') == 'sisu': stats = self.sisu()
elif os.environ.get('SLURM_JOB_PARTITION') == 'halvan': stats = self.halvan()
elif os.environ.get('SNIC_RESOURCE') == 'milou': stats = self.milou()
else: stats = self.local()
# Print the report #
with open(self.p.report, 'w') as handle: handle.write(str(stats))
# Filter short contigs #
FASTA(self.p.Contigs).extract_length(new_path=self.p.filtered, lower_bound=self.length_cutoff)
"betVII-B1",
"betIII-A1",
"betIII-A2",
],
"clades": [
"betIII-A",
"bet-VII-B",
"betVII-A",
"betIV-A",
"Pnec",
"betVI",
"betIII",
"betVII",
"betIV",
"betII",
],
},
"Gammaproteobacteria": {
"species": ["gamII-A1", "gamII-A2", "Acin", "Pseudo-A1", "Pseudo-A2", "Steno"],
"clades": ["gamV-A", "gamIV-A", "gamIII-A", "gamII-A", "gamV", "gamIV", "gamIII", "gamII", "gamI"],
},
"Verrucomicrobia": {"species": ["Xip-A1", "Xip-B1"], "clades": ["verI-B", "verI-A", "verI"]},
"Fibrobacteres": {
"species": ["CLO-84", "Fib", "Cyth", "Ana"],
"clades": ["Anal-A", "Cyth-A", "FibI-A", "OP10I-A", "Anal", "CythI", "Fib-I", "OP10I"],
},
}
species_names = flatter([v["species"] for v in names.values()])
clade_names = flatter([v["clades"] for v in names.values()])
