def init(context, setup, reset, automate):
"""Walk user through setting up a new config file."""
# print a nice welcome message
click.echo(chanjo.__banner__)
cov_tresholds = (context.obj.get('sambamba', {})
.get('cov_treshold', [10, 20]))
defaults = {'sambamba.cov_treshold': {'value': cov_tresholds,
'prompt': 'sufficient coverage'},
'database': {'value': context.obj['database'],
'prompt': 'central database path/URI'}}
if not automate:
questions = [(key, value['prompt'], value['value'])
for key, value in iteritems(defaults)]
# launch init pipeline
init_pipeline('chanjo', context.obj, questions)
else:
logger.info('setting default config values')
for key, value in iteritems(defaults):
context.obj.set(key, value['value'], scope=context.obj.user_data)
# write to the config file
context.obj.save(default_flow_style=False)
if setup:
chanjo_db = Store(uri=context.obj.user_data['database'])
if reset:
chanjo_db.tear_down()
chanjo_db.set_up()
def load_transcripts(sequence,
sample_id=None,
group_id=None,
source=None,
threshold=None):
"""Process a sequence of exon lines.
Args:
sequence (sequence): list of chanjo bed lines
sample_id (Optional[str]): unique sample id, else auto-guessed
grouip_id (Optional[str]): id to group samples
source (Optional[str]): path to coverage source (BAM/Sambamba)
threshold (Optional[int]): completeness level to disqualify exons
Returns:
Result: iterators of `Transcript`, transcripts processed, sample model
"""
exons = sambamba.depth_output(sequence)
transcripts = groupby_tx(exons, sambamba=True)
raw_stats = ((tx_id, tx_stat(tx_id, exons, threshold=threshold))
for tx_id, exons in iteritems(transcripts))
if sample_id is None:
sample_id = next(iter(itervalues(transcripts)))[0]['sampleName']
sample_obj = Sample(id=sample_id, group_id=group_id, source=source)
models = (make_model(sample_obj, tx_id, raw_stat)
for tx_id, raw_stat in raw_stats)
return Result(models=models, count=len(transcripts), sample=sample_obj)
def load_transcripts(sequence, sample_id=None, group_id=None, source=None,
threshold=None):
"""Process a sequence of exon lines.
Args:
sequence (sequence): list of chanjo bed lines
sample_id (Optional[str]): unique sample id, else auto-guessed
grouip_id (Optional[str]): id to group samples
source (Optional[str]): path to coverage source (BAM/Sambamba)
threshold (Optional[int]): completeness level to disqualify exons
Returns:
Result: iterators of `Transcript`, transcripts processed, sample model
"""
exons = sambamba.depth_output(sequence)
transcripts = groupby_tx(exons, sambamba=True)
raw_stats = ((tx_id, tx_stat(tx_id, exons, threshold=threshold))
for tx_id, exons in iteritems(transcripts))
if sample_id is None:
sample_id = next(iter(itervalues(transcripts)))[0]['sampleName']
sample_obj = Sample(id=sample_id, group_id=group_id, source=source)
models = (make_model(sample_obj, tx_id, raw_stat) for tx_id, raw_stat
in raw_stats)
return Result(models=models, count=len(transcripts), sample=sample_obj)
def expand_row(header, row):
"""Parse information in row to dict.
Args:
header (dict): key/index header dict
row (List[str]): sambamba BED row
Returns:
dict: parsed sambamba output row
"""
thresholds = {
threshold: float(row[key])
for threshold, key in iteritems(header['thresholds'])
}
data = {
'chrom': row[0],
'chromStart': int(row[1]),
'chromEnd': int(row[2]),
'sampleName': row[header['sampleName']],
'readCount': int(row[header['readCount']]),
'meanCoverage': float(row[header['meanCoverage']]),
'thresholds': thresholds,
'extraFields': row[header['extraFields']]
}
return data
def init(context, setup, reset, automate, transcripts):
"""Walk user through setting up a new config file."""
# print a nice welcome message
click.echo(chanjo.__banner__)
cov_tresholds = (context.obj.get('sambamba',
{}).get('cov_treshold', [10, 20]))
defaults = {
'sambamba.cov_treshold': {
'value': cov_tresholds,
'prompt': 'sufficient coverage'
},
'database': {
'value': str(context.obj['database']),
'prompt': 'central database path/URI'
}
}
if not automate:
questions = [(key, value['prompt'], value['value'])
for key, value in iteritems(defaults)]
# launch init pipeline
init_pipeline('chanjo', context.obj, questions)
else:
logger.info('setting default config values')
for key, value in iteritems(defaults):
context.obj.set(key, value['value'], scope=context.obj.user_data)
# write to the config file
context.obj.save(default_flow_style=False)
if setup:
only_tx = transcripts or context.obj.get('transcripts') or False
base = TXBASE if only_tx else BASE
chanjo_db = Store(uri=context.obj['database'], base=base)
if reset:
chanjo_db.tear_down()
chanjo_db.set_up()
def link_elements(sequence):
"""Process a sequence of exon lines.
Args:
sequence (sequence): list of chanjo bed lines
Returns:
Result: iterators of transcript models, number of transcripts processed
"""
exons = parse_bed.chanjo(sequence)
transcripts = groupby_tx(exons)
models = (make_model(tx_id, exons) for tx_id, exons in
iteritems(transcripts))
return Result(models=models, count=len(transcripts))
def tx_stat(transcript_id, exons, threshold=None):
"""Calculate metrics for transcript stats model.
Args:
transcript_id (str): unqiue transcript id
exons (List[dict]): list of exon transcripts
threshold (Optional[int]): completeness level to disqualify exons
Returns:
dict: aggregated stats over all exons
"""
sums = {'bases': 0, 'mean_coverage': 0}
incomplete_exons = []
# for each of the exons (linked to one transcript)
for exon in exons:
# go over each of the fields to sum up
exon_length = (exon['chromEnd'] - exon['chromStart'])
sums['bases'] += exon_length
sums['mean_coverage'] += (exon['meanCoverage'] * exon_length)
# add to the total sum for completeness levels
for comp_key in [10, 15, 20, 50, 100]:
if comp_key in exon['thresholds']:
sums_key = "completeness_{}".format(comp_key)
if sums_key not in sums:
sums[sums_key] = 0
completeness = exon['thresholds'][comp_key]
sums[sums_key] += (completeness * exon_length)
if threshold == comp_key and completeness < 100:
exon_obj = Exon(exon['chrom'], exon['chromStart'],
exon['chromEnd'], completeness)
incomplete_exons.append(exon_obj)
fields = {
key: (value / sums['bases'])
for key, value in iteritems(sums) if key != 'bases'
}
fields['incomplete_exons'] = incomplete_exons
fields['threshold'] = threshold
return fields
def statistics(data, sample_obj, exon_obj):
"""Create models from a sambamba output row.
Args:
data (dict): parsed sambamba output row
sample_obj (Sample): linked sample model
exon_obj (Exon): linked exon model
Returns:
List[ExonStatistic]: stats models linked to exon and sample
"""
relationships = dict(sample=sample_obj, exon=exon_obj)
stats = [ExonStatistic(metric='mean_coverage', value=data['meanCoverage'],
**relationships)]
for threshold, value in iteritems(data['thresholds']):
metric = "completeness_{}".format(threshold)
stat = ExonStatistic(metric=metric, value=value, **relationships)
stats.append(stat)
return stats
def tx_stat(transcript_id, exons, threshold=None):
"""Calculate metrics for transcript stats model.
Args:
transcript_id (str): unqiue transcript id
exons (List[dict]): list of exon transcripts
threshold (Optional[int]): completeness level to disqualify exons
Returns:
dict: aggregated stats over all exons
"""
sums = {'bases': 0, 'mean_coverage': 0}
incomplete_exons = []
# for each of the exons (linked to one transcript)
for exon in exons:
# go over each of the fields to sum up
exon_length = (exon['chromEnd'] - exon['chromStart'])
sums['bases'] += exon_length
sums['mean_coverage'] += (exon['meanCoverage'] * exon_length)
# add to the total sum for completeness levels
for comp_key in [10, 15, 20, 50, 100]:
if comp_key in exon['thresholds']:
sums_key = "completeness_{}".format(comp_key)
if sums_key not in sums:
sums[sums_key] = 0
completeness = exon['thresholds'][comp_key]
sums[sums_key] += (completeness * exon_length)
if threshold == comp_key and completeness < 100:
exon_obj = Exon(exon['chrom'], exon['chromStart'],
exon['chromEnd'], completeness)
incomplete_exons.append(exon_obj)
fields = {key: (value / sums['bases']) for key, value in iteritems(sums)
if key != 'bases'}
fields['incomplete_exons'] = incomplete_exons
fields['threshold'] = threshold
return fields
def expand_row(header, row):
"""Parse information in row to dict.
Args:
header (dict): key/index header dict
row (List[str]): sambamba BED row
Returns:
dict: parsed sambamba output row
"""
thresholds = {threshold: float(row[key])
for threshold, key in iteritems(header['thresholds'])}
data = {
'chrom': row[0],
'chromStart': int(row[1]),
'chromEnd': int(row[2]),
'sampleName': row[header['sampleName']],
'readCount': int(row[header['readCount']]),
'meanCoverage': float(row[header['meanCoverage']]),
'thresholds': thresholds,
'extraFields': row[header['extraFields']]
}
return data
def statistics(data, sample_obj, exon_obj=None, exon_id=None):
"""Create models from a sambamba output row.
Args:
data (dict): parsed sambamba output row
sample_obj (Sample): linked sample model
exon_obj (int): primary key for related Exon
Returns:
List[ExonStatistic]: stats models linked to exon and sample
"""
if exon_obj:
relationships = dict(sample=sample_obj, exon=exon_obj)
else:
relationships = dict(sample=sample_obj, exon_id=exon_id)
stats = [ExonStatistic(metric='mean_coverage', value=data['meanCoverage'],
**relationships)]
for threshold, value in iteritems(data['thresholds']):
metric = "completeness_{}".format(threshold)
stat = ExonStatistic(metric=metric, value=value, **relationships)
stats.append(stat)
return stats
