def fill_small_wholes(self,
max_size,
write_holes_to_file=None,
touched_nodes=None):
cleaner = HolesCleaner(self, max_size, touched_nodes=touched_nodes)
areas = cleaner.run()
n_filled = 0
hole_intervals = []
for node_id in areas.areas:
if touched_nodes is not None:
if node_id not in touched_nodes:
continue
starts = areas.get_starts(node_id)
ends = areas.get_ends(node_id)
for start, end in zip(starts, ends):
self.data[node_id].set_interval_value(start, end, True)
logging.debug("Filling hole %s, %d, %d" %
(node_id, start, end))
n_filled += 1
assert end - start <= max_size
hole_intervals.append(Interval(start, end, [node_id]))
logging.info("Filled %d small holes (splitted into holes per node)" %
n_filled)
if write_holes_to_file is not None:
intervals = IntervalCollection(hole_intervals)
intervals.to_file(write_holes_to_file, text_file=True)
self.sanitize()
def check_similarity(self, analyse_first_n_peaks=10000000):
print("Number of peaks in main set: %d" % len(self.peaks1.intervals))
self.results.tot_peaks1 = len(self.peaks1.intervals)
self.results.tot_peaks2 = len(self.peaks2.intervals)
counter = 0
visited = set([])
for peak in sorted(self.peaks1, key=lambda x: x.score,
reverse=True)[0:analyse_first_n_peaks]:
assert peak.unique_id is not None
counter += 1
if counter % 500 == 0:
logging.info("Checked %d peaks" % counter)
touching = self.peaks2.approx_contains_part_of_interval(
peak, visited)
if touching:
visited.add(touching[0].unique_id)
self.peaks2_in_peaks1.append(touching[0])
self.peaks1_in_peaks2.append(peak)
else:
self.peaks1_not_in_peaks2.append(peak)
for peak in self.peaks2:
if peak.unique_id not in visited:
self.peaks2_not_in_peaks1.append(peak)
self.results.peaks1_in_peaks2 = len(self.peaks1_in_peaks2)
self.results.peaks2_in_peaks1 = len(self.peaks2_in_peaks1)
self.results.peaks1_not_in_peaks2 = len(self.peaks1_not_in_peaks2)
self.results.peaks2_not_in_peaks1 = len(self.peaks2_not_in_peaks1)
chromosome = self.chromosome
if chromosome is None:
chromosome = "unknown"
gpc_not_matching_macs = IntervalCollection(self.peaks1_not_in_peaks2)
gpc_not_matching_macs.to_file("gpc_not_matching_macs_chr%s.intervals" %
chromosome,
text_file=True)
logging.info(
"Wrote peaks not matching to file gpc_not_matching_macs_chr%s.intervals"
% chromosome)
macs_not_matching_gpc = IntervalCollection(self.peaks2_not_in_peaks1)
macs_not_matching_gpc.to_file("macs_not_matching_gpc_chr%s.intervals" %
chromosome,
text_file=True)
logging.info(
"Wrote peaks not matching to file macs_not_matching_gpc_chr%s.intervals"
% chromosome)
def find_linear_path_through_chromosome(chromosome, chromend, fasta_file_name,
ob_graph_file_name,
vg_graph_file_name):
genome = Fasta(fasta_file_name)
seq = str(genome[chromosome][0:50818468]).lower()
logging.info("Creating sequence retriever")
sequence_retriever = SequenceRetriever.from_vg_json_graph(
vg_graph_file_name)
graph = GraphWithReversals.from_numpy_file(ob_graph_file_name)
start_nodes = graph.get_first_blocks()
assert len(start_nodes) == 1, "Found %d start nodes" % start_nodes
start_node = start_nodes[0]
traverser = GraphTraverserUsingSequence(graph, seq, sequence_retriever)
traverser.search_from_node(start_node)
path = traverser.get_interval_found()
path = IntervalCollection(path)
path.to_file("22_path.intervalcollection", text_file=True)
logging.info("Done")
def check_similarity_old(self, analyse_first_n_peaks=10000000):
i = 1
for peak_datasets in [(self.peaks1, self.peaks2),
(self.peaks2, self.peaks1)]:
n_identical = 0
tot_n_similar = 0
n_similar = 0
n_tot = 0
print("\n-- Comparing set %d against set %d ---" % (i, i % 2 + 1))
peaks1, peaks2 = peak_datasets
print("Number of peaks in main set: %d" % len(peaks1.intervals))
if i == 1:
self.results.tot_peaks1 = len(peaks1.intervals)
else:
self.results.tot_peaks2 = len(peaks1.intervals)
not_matching = []
matching = []
counter = 0
visited = set([])
for peak in sorted(peaks1, key=lambda x: x.score,
reverse=True)[0:analyse_first_n_peaks]:
assert peak.unique_id is not None
counter += 1
if counter % 500 == 0:
logging.info("Checked %d peaks" % counter)
touching = peaks2.approx_contains_part_of_interval(
peak, visited)
if touching:
visited.add(touching[0].unique_id)
n_similar += 1
if i == 1:
self.peaks1_in_peaks2.append(peak)
else:
self.peaks2_in_peaks1.append(peak)
matching.append(peak)
else:
not_matching.append(peak)
if i == 1:
self.peaks1_not_in_peaks2.append(peak)
else:
self.peaks2_not_in_peaks1.append(peak)
n_tot += 1
self.results.peaks1_in_peaks2 = len(self.peaks1_in_peaks2)
self.results.peaks2_in_peaks1 = len(self.peaks2_in_peaks1)
self.results.peaks1_not_in_peaks2 = len(self.peaks1_not_in_peaks2)
self.results.peaks2_not_in_peaks1 = len(self.peaks2_not_in_peaks1)
not_matching = IntervalCollection(not_matching)
not_matching.to_file("not_matching_set%d.intervals" % i,
text_file=True)
logging.info(
"Wrote peaks not matching to file not_matching_set%d.intervals"
% i)
matching = IntervalCollection(matching)
matching.to_file("matching_set%d.intervals" % i, text_file=True)
logging.info("Total peaks in main set: %d" % n_tot)
logging.info("N similar to peak in other set: %d " % n_similar)
logging.info("N not matching other set: %d " %
len(not_matching.intervals))
i += 1