def call_debruijn_graph(self, windows, reads):
"""Helper function to call debruijn_graph module."""
windows_haplotypes = []
# Build and process de-Bruijn graph for each window.
sam_reader = sam.InMemorySamReader(reads)
for window in windows:
if window.end - window.start > self.config.ws_config.max_window_size:
continue
if not self.ref_reader.is_valid(window):
continue
ref = self.ref_reader.query(window)
window_reads = list(sam_reader.query(window))
with timer.Timer() as t:
graph = debruijn_graph.build(ref, window_reads, self.config.dbg_config)
graph_building_time = t.GetDuration()
if not graph:
candidate_haplotypes = [ref]
else:
candidate_haplotypes = graph.candidate_haplotypes()
if candidate_haplotypes and candidate_haplotypes != [ref]:
candidate_haplotypes_info = realigner_pb2.CandidateHaplotypes(
span=window, haplotypes=candidate_haplotypes)
windows_haplotypes.append(candidate_haplotypes_info)
self.diagnostic_logger.log_graph_metrics(
window, graph, candidate_haplotypes, graph_building_time)
return windows_haplotypes
def _initialize(self):
"""Initialize the resources needed for this work in the current env."""
if self.initialized:
raise ValueError('Cannot initialize this object twice')
self.ref_reader = fasta.IndexedFastaReader(
self.options.reference_filename)
self.sam_reader = self._make_sam_reader()
self.in_memory_sam_reader = sam.InMemorySamReader([])
if self.options.realigner_enabled:
self.realigner = realigner.Realigner(
self.options.realigner_options, self.ref_reader)
self.pic = pileup_image.PileupImageCreator(
ref_reader=self.ref_reader,
sam_reader=self.in_memory_sam_reader,
options=self.options.pic_options)
if in_training_mode(self.options):
self.labeler = self._make_labeler_from_options()
self.variant_caller = variant_caller.VariantCaller(
self.options.variant_caller_options)
self.random = np.random.RandomState(self.options.random_seed)
self.initialized = True
