def profile(self):
"""Big deal function"""
# So we start with essential stats. In the section below, we will simply go through each contig
# in the BAM file and populate the contigs dictionary for the first time.
for i in range(0, len(self.contig_names)):
contig_name = self.contig_names[i]
contig = contigops.Contig(contig_name)
contig.length = self.contig_lengths[i]
contig.split_length = self.a_meta['split_length']
contig.min_coverage_for_variability = self.min_coverage_for_variability
contig.skip_SNV_profiling = self.skip_SNV_profiling
contig.report_variability_full = self.report_variability_full
self.progress.new('Profiling "%s" (%d of %d) (%s nts)' %
(contig.name, i + 1, len(
self.contig_names), pp(int(contig.length))))
# populate contig with empty split objects and
for split_name in self.contig_name_to_splits[contig_name]:
s = self.splits_basic_info[split_name]
split_sequence = self.contig_sequences[contig_name][
'sequence'][s['start']:s['end']]
split = contigops.Split(split_name, split_sequence,
contig_name, s['order_in_parent'],
s['start'], s['end'])
contig.splits.append(split)
# analyze coverage for each split
contig.analyze_coverage(self.bam, self.progress)
# test the mean coverage of the contig.
discarded_contigs_due_to_C = set([])
if contig.coverage.mean < self.min_mean_coverage:
# discard this contig and continue
discarded_contigs_due_to_C.add(contig.name)
self.progress.end()
continue
if not self.skip_SNV_profiling:
contig.analyze_auxiliary(self.bam, self.progress)
self.progress.end()
# add contig to the dict.
self.contigs[contig_name] = contig
if discarded_contigs_due_to_C:
self.run.info('contigs_after_C', pp(len(self.contigs)))
# set contig abundance
contigops.set_contigs_abundance(self.contigs)
self.check_contigs()
def profile_contig_worker(available_index_queue, output_queue, info_dict):
bam_file = pysam.Samfile(info_dict['input_file_path'], 'rb')
while True:
index = available_index_queue.get(True)
contig_name = info_dict['contig_names'][index]
contig = contigops.Contig(contig_name)
contig.length = info_dict['contig_lengths'][index]
contig.split_length = info_dict['split_length']
contig.min_coverage_for_variability = info_dict[
'min_coverage_for_variability']
contig.skip_SNV_profiling = info_dict['skip_SNV_profiling']
contig.report_variability_full = info_dict[
'report_variability_full']
# populate contig with empty split objects and
for split_name in info_dict['contig_name_to_splits'][contig_name]:
s = info_dict['splits_basic_info'][split_name]
split_sequence = info_dict['contig_sequences'][contig_name][
'sequence'][s['start']:s['end']]
split = contigops.Split(split_name, split_sequence,
contig_name, s['order_in_parent'],
s['start'], s['end'])
contig.splits.append(split)
# analyze coverage for each split
contig.analyze_coverage(bam_file)
# test the mean coverage of the contig.
if contig.coverage.mean < info_dict['min_mean_coverage']:
output_queue.put(None)
continue
if not info_dict['skip_SNV_profiling']:
contig.analyze_auxiliary(bam_file)
output_queue.put(contig)
for split in contig.splits:
del split.coverage
del split.auxiliary
del split
del contig.splits[:]
del contig.coverage
del contig
# we are closing this object here for clarity, although w
# are not really closing it since the code never reaches here
# and the worker is killed by its parent:
bam_file.close()
return
def profile_contig_worker(self, available_index_queue, output_queue):
bam_file = pysam.Samfile(self.input_file_path, 'rb')
while True:
index = available_index_queue.get(True)
contig_name = self.contig_names[index]
contig = contigops.Contig(contig_name)
contig.length = self.contig_lengths[index]
contig.split_length = self.a_meta['split_length']
contig.min_coverage_for_variability = self.min_coverage_for_variability
contig.skip_SNV_profiling = self.skip_SNV_profiling
contig.report_variability_full = self.report_variability_full
contig.ignore_orphans = not self.include_orphans
contig.max_coverage_depth = self.max_coverage_depth
# populate contig with empty split objects and
for split_name in self.contig_name_to_splits[contig_name]:
s = self.splits_basic_info[split_name]
split_sequence = self.contig_sequences[contig_name][
'sequence'][s['start']:s['end']]
split = contigops.Split(split_name, split_sequence,
contig_name, s['order_in_parent'],
s['start'], s['end'])
contig.splits.append(split)
# analyze coverage for each split
contig.analyze_coverage(bam_file)
# test the mean coverage of the contig.
if contig.coverage.mean < self.min_mean_coverage:
output_queue.put(None)
continue
if not self.skip_SNV_profiling:
contig.analyze_auxiliary(bam_file)
codons_in_genes_to_profile_SCVs = set([])
for split in contig.splits:
for column_profile in list(split.column_profiles.values()):
pos_in_contig = column_profile['pos_in_contig']
column_profile['in_partial_gene_call'], \
column_profile['in_complete_gene_call'],\
column_profile['base_pos_in_codon'] = self.get_nt_position_info(contig.name, pos_in_contig)
column_profile['sample_id'] = self.sample_id
column_profile[
'corresponding_gene_call'] = -1 # this means there is no gene call that corresponds to this
# nt position, which will be updated in the following lines.
# yeah, we use '-1', because genecaller ids start from 0 :/
column_profile['codon_order_in_gene'] = -1
# if this particular position (`pos_in_contig`) falls within a COMPLETE gene call,
# we would like to find out which unique gene caller id(s) match to this position.
if column_profile['in_complete_gene_call']:
corresponding_gene_caller_ids = self.get_corresponding_gene_caller_ids_for_base_position(
contig.name, pos_in_contig)
# if there are more than one corresponding gene call, this usually indicates an assembly error
# just to be on the safe side, we will not report a corresopnding unique gene callers id for this
# position
if len(corresponding_gene_caller_ids) == 1:
# if we are here, it means this nucleotide position is in a complete gene call. we will do two things here.
# first, we will store the gene_callers_id that corresponds to this nt position, and then we will store the
# order of the corresponding codon in the gene for this nt position.
gene_callers_id = corresponding_gene_caller_ids[
0]
column_profile[
'corresponding_gene_call'] = gene_callers_id
column_profile[
'codon_order_in_gene'] = self.get_corresponding_codon_order_in_gene(
gene_callers_id, contig.name,
pos_in_contig)
# save this information for later use
codons_in_genes_to_profile_SCVs.add(
(gene_callers_id,
column_profile['codon_order_in_gene']), )
codon_frequencies = bamops.CodonFrequencies()
codons_in_genes_to_profile_SCVs_dict = {}
for gene_callers_id, codon_order in codons_in_genes_to_profile_SCVs:
if gene_callers_id not in codons_in_genes_to_profile_SCVs_dict:
codons_in_genes_to_profile_SCVs_dict[
gene_callers_id] = set([])
codons_in_genes_to_profile_SCVs_dict[gene_callers_id].add(
codon_order)
gene_caller_ids_to_profile = list(
codons_in_genes_to_profile_SCVs_dict.keys())
for i in range(len(gene_caller_ids_to_profile)):
gene_callers_id = gene_caller_ids_to_profile[i]
codons_to_profile = codons_in_genes_to_profile_SCVs_dict[
gene_callers_id]
gene_call = self.genes_in_contigs_dict[gene_callers_id]
contig_name = gene_call['contig']
contig.codon_frequencies_dict[
gene_callers_id] = codon_frequencies.process_gene_call(
bam_file, gene_call,
self.contig_sequences[contig_name]['sequence'],
codons_to_profile)
output_queue.put(contig)
for split in contig.splits:
del split.coverage
del split.auxiliary
del split
del contig.splits[:]
del contig.coverage
del contig
# we are closing this object here for clarity, although w
# are not really closing it since the code never reaches here
# and the worker is killed by its parent:
bam_file.close()
return
