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 get_intersecting_intervals(args):
from offsetbasedgraph import IntervalCollection
intervals1 = IntervalCollection.from_file(args.file1, text_file=True, graph=args.graph)
intervals2 = IntervalCollection.from_file(args.file2, text_file=True, graph=args.graph)
out = []
for interval1 in intervals1.intervals:
for interval2 in intervals2.intervals:
if interval1.intersects(interval2):
out.append(interval1)
logging.info("Found match between %s and %s" % (interval1, interval2))
continue
IntervalCollection(out).to_file(args.out_file_name, text_file=True)
logging.info("Wrote intersecting intervals to %s" % args.out_file_name)
def create_linear_peaks_from_bed(linear_sequence_fasta_file, peaks_bed_file,
obg_graph_file_name, vg_graph_file_name,
start_node, region):
ob_graph = obg.GraphWithReversals.from_file(obg_graph_file_name)
search_sequence = open(linear_sequence_fasta_file).read()
sequence_retriever = SequenceRetriever.from_vg_graph(vg_graph_file_name)
traverser = GraphTraverserUsingSequence(ob_graph, search_sequence,
sequence_retriever)
traverser.search_from_node(start_node)
linear_path_interval = traverser.get_interval_found()
IntervalCollection([linear_path_interval
]).to_file("linear_path.intervalcollection",
text_file=True)
print("Length")
print(linear_path_interval.length())
print(linear_path_interval.region_paths[0])
print(linear_path_interval.start_position)
print(linear_path_interval.end_position)
linear_peaks = PeakCollection.create_from_linear_intervals_in_bed_file(
obg_graph_file_name, linear_path_interval, peaks_bed_file,
region.start, region.end)
linear_peaks.to_file("linear_peaks.intervalcollection", text_file=True)
def vg_path_to_obg_interval(path_file_name, out_file_name):
json_objects = get_json_lines(path_file_name)
alignments = (Alignment.from_json(json_object)
for json_object in json_objects)
intervals = []
for alignment in alignments:
path = alignment.path
interval = path.to_obg()
intervals.append(interval)
chrom = path.name
start_node = path.mappings[0].node_id(
) # [m.node_id() for m in path.mappings]
logging.info("Processing chromosome %s with start node %d" %
(chrom, start_node))
with open("chr%s_start_node.txt" % chrom, "w") as f:
f.write(str(start_node))
file_name = out_file_name.split(
".")[0] + "_" + chrom + "." + out_file_name.split(".")[-1]
IntervalCollection([interval]).to_file(file_name, text_file=True)
logging.info("Number of files in interval for chrom %s: %d" %
(chrom, len(interval.region_paths)))
logging.info("Wrote path as obg interval to %s" % file_name)
def call_peaks(self):
genome_size = sum(block.length()
for block in self.graph.blocks.values())
experiment_info = ExperimentInfo(genome_size, 50, 20)
experiment_info.n_sample_reads = self.n_sample_reads
experiment_info.n_control_reads = self.n_control_reads
snarlbuilder = SnarlGraphBuilder(self.graph.copy(),
self.snarls,
id_counter=self.graph.max_block_id() +
1)
snarlgraph = snarlbuilder.build_snarl_graphs()
linear_map = LinearSnarlMap(snarlgraph, self.graph)
linear_map.to_file("simulated_snarl_map.tmp")
caller = CallPeaks(self.graph,
sample_intervals="dummy",
control_intervals=IntervalCollection(
self.control_reads),
experiment_info=experiment_info,
has_control=self.with_control,
linear_map="simulated_snarl_map.tmp")
caller._sample_pileup = self.sample_pileup
self.sample_pileup.to_bed_graph("sample.bdg")
caller.create_control(True)
caller.scale_tracks(True)
caller.get_score()
caller.call_peaks()
sequence_retriever = DummySequenceRetriever()
caller.save_max_path_sequences_to_fasta_file(
"simulated_peak_sequences.fasta", sequence_retriever)
self.caller = caller
def assert_final_peaks_equals_input_peaks(self):
final_peaks = IntervalCollection.create_list_from_file(
"test_max_paths.intervalcollection")
for peak in self.peaks:
self.assertTrue(
peak in final_peaks.intervals,
"Peak %s not in final peaks. Final peaks: \n%s" %
(peak, final_peaks.intervals))
self.assertEqual(len(self.peaks), len(final_peaks.intervals))
def from_file(cls, file_name, graph):
logging.info("Reading from file")
logging.info("Reading dict structure")
with open(file_name, "rb") as f:
node_dict = pickle.load(f)
logging.info("Reading intervals")
intervals = IntervalCollection.from_file(file_name + ".intervals",
graph=graph)
return cls(node_dict, graph, list(intervals))
def _read_alignments(self):
if self.alignment_file_name.endswith(".json"):
self.alignments = vg_json_file_to_interval_collection(self.alignment_file_name).intervals
elif self.alignment_file_name.endswith(".graphnodes"):
self.alignments = (Interval(0, 1, [int(n) for n in line.strip().split()[1].split(",")])
for line in open(self.alignment_file_name))
elif self.alignment_file_name.endswith(".graphalignments"):
self.alignments = (Interval.from_file_line(line.strip().split("\t")[1]) for line in open(self.alignment_file_name) if line.strip().split("\t")[1] != ".")
else:
self.alignments = IntervalCollection.from_file(self.alignment_file_name).intervals
def test_count_unique_reads(self):
reads = [
IntervalCollection([
Interval(4, 10, [1, 2, 3]),
Interval(4, 5, [1]),
Interval(5, 5, [1]),
Interval(6, 2, [-3, -2, -1])
])
]
unique = MultipleGraphsCallpeaks.count_number_of_unique_reads(reads)
self.assertEqual(unique, 3)
def run_predict_path(args):
chromosomes = args.chromosomes.split(",")
processes = []
if not os.path.isfile(args.alignments):
logging.error("Input alignments file %s does not exist" %
args.alignments)
sys.exit()
for chromosome in chromosomes:
logging.info("Starting process for chromosome %s " % chromosome)
process = Process(target=run_predict_path_single_chromosome,
args=(args.alignments, chromosome, args.data_dir,
args.linear_ref_bonus, args.out_file_name,
args.max_nodes_to_traverse))
process.start()
processes.append(process)
for process in processes:
process.join()
# Merge all fasta files that were produces
out_fasta = open(args.out_file_name + ".fa", "w")
logging.info("Merging fasta files")
for chromosome in tqdm(chromosomes):
with open(args.out_file_name + "_" + chromosome + ".fasta") as f:
out_fasta.write(f.read())
logging.info("Wrote resulting linear reference to %s" %
(args.out_file_name + ".fa"))
# Create indexed intervals for each interval file that was produced
logging.info("Creating indexed interval for all chromosomes")
for chromosome in chromosomes:
file_name = args.out_file_name + "_" + chromosome + ".intervalcollection"
graph = Graph.from_file(args.data_dir + chromosome + ".nobg")
intervals = IntervalCollection.from_file(file_name,
text_file=True,
graph=graph)
intervals = list(intervals.intervals)
assert len(
intervals) == 1, "Only a single interval in file is supported"
interval = intervals[0]
indexed = interval.to_numpy_indexed_interval()
indexed.to_file(file_name + ".indexed")
logging.info("Wrote indexed interval to file %s" % file_name +
".indexed")
if not args.skip_bwa_index:
logging.info("Running bwa index")
run_bwa_index(args.out_file_name + ".fa")
else:
logging.info("Not creating bwa index")
def test_filter_duplicates(self):
intervals = [
Interval(0, 10, [1, 2, 3]),
Interval(1, 10, [1, 2, 3]),
Interval(0, 10, [1, 2, 3])
]
interval_collection = IntervalCollection(intervals)
intervals_filtered = list(UniqueIntervals(interval_collection))
self.assertEqual(len(intervals_filtered), len(intervals) - 1)
self.assertEqual(intervals_filtered[0], intervals[0])
self.assertEqual(intervals_filtered[1], intervals[1])
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 __init__(self, graph, sequence_retriever, linear_path_file_name,
peaks1_file_name, peaks2_file_name):
self.graph = graph
self.sequence_retriever = sequence_retriever
self.peaks1 = PeakCollection.create_list_from_file(peaks1_file_name,
graph=graph)
self.peaks2 = PeakCollection.create_list_from_file(peaks2_file_name,
graph=graph)
print("Number of intervals in set 1/2: %d / %d" %
(len(self.peaks1.intervals), len(self.peaks2.intervals)))
if linear_path_file_name is not None:
self.linear_path = IntervalCollection.create_list_from_file(
linear_path_file_name, self.graph).intervals[0]
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 test_complex_graph(self):
intervals = IntervalCollection([
Interval(0, 3, [1, 3, 4, 6, 10]),
Interval(1, 2, [2]),
Interval(2, 3, [2]),
Interval(0, 3, [7, 9])
])
haplotyper = HaploTyper(self.complex_graph, intervals)
haplotyper.build()
max_interval = haplotyper.get_maximum_interval_through_graph()
self.assertEqual(
max_interval,
Interval(0, 3, [1, 2, 7, 9, 10, 12])
)
def test_all_steps(self):
run_argument_parser([
"create_ob_graph", "-o", "tests/testgraph.obg",
"tests/vg_test_graph.json"
])
run_argument_parser(
['create_linear_map', "--graph", "tests/testgraph.obg"])
IntervalCollection([Interval(1, 1, [1, 2])
]).to_file("tests/sample.intervalcollection")
run_argument_parser([
"callpeaks", "--graph", "tests/testgraph.obg", "-s",
"tests/sample.intervalcollection", "-n", "tests/test_experiment_",
"-f", "10", "-r", "7"
])
def analyse_pileups_on_peaks(ob_graph, pileups_file_names,
peak_intervals_file_name):
print("Analysing peaks")
pileups = {
name: SparsePileup.from_bed_graph(ob_graph, pileup)
for name, pileup in pileups_file_names.items()
}
peaks = IntervalCollection.from_file(peak_intervals_file_name,
text_file=True)
for peak in peaks:
print()
print("Peak %s" % peak)
rp = peak.region_paths[0]
for name, pileup in pileups.items():
pileup_sum = sum(pileup.data[rp].sum() for rp in peak.region_paths)
print("Pileup %s: %d" % (name, pileup_sum))
def make_haplotype_fasta(chromosome, haplotype, data_dir):
s = SequenceGraph.from_file(data_dir + "giab_chr" + chromosome +
".nobg.sequences")
print("Getting interval")
interval = list(
IntervalCollection.from_file(data_dir + "haplotype_" + chromosome +
"__" + haplotype + ".intervalcollection",
text_file=True).intervals)[0]
print("Getting sequence")
sequence = s.get_interval_sequence(interval)
print("Writing to file")
f = open(
data_dir + "giab_chr" + chromosome + "_haplotype" + haplotype +
".fasta", "w")
f.write(">seq\n%s\n" % sequence)
f.close()
f.close()
def test_simple(self):
graph = Graph(
{i: Block(3) for i in range(1, 5)},
{
1: [2, 3],
2: [4],
3: [4]
}
)
graph.convert_to_numpy_backend()
intervals = IntervalCollection([
Interval(0, 3, [1, 3])
])
haplotyper = HaploTyper(graph, intervals)
haplotyper.build()
max_interval = haplotyper.get_maximum_interval_through_graph()
self.assertEqual(
max_interval,
Interval(0, 3, [1, 3, 4])
)
def vg_alignments_to_linear():
ob_graph = obg.GraphWithReversals.from_file("haplo1kg50-mhc.obg")
vg_graph = pyvg.vg.Graph.create_from_file("haplo1kg50-mhc.json")
path = create_linear_path(ob_graph, vg_graph)
analyser = AlignmentsAnalyser(
vg_graph, "ENCFF001HNI_haplo1kg50-mhc_filtered_q50.gam", ob_graph,
path) # sample reads
#linear = analyser.to_linear_alignments()
#collection = IntervalCollection(linear)
#collection.to_file("graph_reads_on_linear2.intervals")
linear = IntervalCollection.from_file(
"graph_reads_on_linear2.intervals").intervals
#linear = IntervalCollection.create_list_from_file("graph_reads_on_linear.intervals")
f = open("graph_reads_on_linear.bed", "w")
path = path.to_indexed_interval()
linear_reads = []
for read in linear:
read.graph = ob_graph
assert np.all(np.array(read.region_paths) > 0) or np.all(
np.array(read.region_paths) < 0)
dir = "+"
if read.region_paths[0] < 0:
dir = "-"
read = read.get_reverse()
graph_start = read.start_position
graph_end = read.end_position
linear_start = MHC_REGION.start + path.get_offset_at_position(
graph_start)
linear_end = MHC_REGION.start + path.get_offset_at_position(graph_end)
f.writelines("chr6\t%d\t%d\t.\t0\t%s\n" %
(linear_start, linear_end, dir))
f.close()
def to_file(self, file_name):
logging.info("Writing to file")
with open(file_name, "wb") as f:
pickle.dump(self._node_dict, f)
IntervalCollection(self.intervals).to_file(file_name + ".intervals")
def make_haplotype_paths(graph_file_name, linear_ref_path_file_name,
haplotype0_file_name, haplotype1_file_name,
out_base_name, chromosome):
# Make a linear reference fasta and interval and haplotypes fasta and intervals
chrom = chromosome
graph = Graph.from_file(graph_file_name)
sequence_graph = SequenceGraph.from_file(graph_file_name + ".sequences")
linear_ref = IntervalCollection.from_file(linear_ref_path_file_name,
text_file=True)
linear_ref = list(linear_ref.intervals)[0]
linear_ref_nodes = set(linear_ref.region_paths)
# Write linear ref fasta to file
linear_ref_seq = sequence_graph.get_interval_sequence(linear_ref)
out_file = open("linear_ref_" + chrom + ".fasta", "w")
out_file.writelines([">%s\n" % chrom])
out_file.writelines([linear_ref_seq + "\n"])
out_file.close()
logging.info("Wrote linear ref sequence. N nodes in linear ref: %d" %
len(linear_ref_nodes))
haplotype_nodes = [set(), set()] # For haplotype 0 and 1
for haplotype in [0, 1]:
haplotype_file_name = haplotype0_file_name
if haplotype == 1:
haplotype_file_name = haplotype1_file_name
intervals = vg_json_file_to_intervals(haplotype_file_name, graph)
for interval in intervals:
for node in interval.region_paths:
haplotype_nodes[haplotype].add(node)
logging.info("N nodes in haplotype 0: %d" % len(haplotype_nodes[0]))
logging.info("N nodes in haplotype 0 that are also in linear ref: %d" %
len(haplotype_nodes[0].intersection(linear_ref_nodes)))
logging.info("N nodes in haplotype 1: %d" % len(haplotype_nodes[1]))
# Traverse graph to get full correct haplotype intervals
first_nodes = graph.get_first_blocks()
assert len(first_nodes) == 1
logging.info("N nodes in graph: %d" % len(graph.blocks))
for haplotype in [0, 1]:
logging.info("Traversing haplotype %d" % haplotype)
nodes = []
node = first_nodes[0]
nodes_in_haplotype = haplotype_nodes[haplotype]
nodes_in_haplotype = set(range(
0, max(linear_ref_nodes))).difference(linear_ref_nodes)
logging.info("There are %d haplotype nodes" % len(nodes_in_haplotype))
assert len(
nodes_in_haplotype
) > 0, "There are no haplotype nodes. Check that haplotype json files are not empty"
n_haplotype_nodes = 0
i = 0
while True:
nodes.append(node)
if i % 50000 == 0:
logging.info("#%d nodes traversed. On node %d" % (i, node))
i += 1
next_nodes = set(graph.adj_list[node])
if len(next_nodes) == 0:
logging.info("Reached end node %d with 0 edges" % node)
break
next_on_haplotype = next_nodes.intersection(nodes_in_haplotype)
if len(next_on_haplotype) == 1:
n_haplotype_nodes += 1
next_node = list(next_on_haplotype)[0]
assert next_node != node
node = next_node
elif len(next_on_haplotype) == 0:
logging.debug(
"No new haplotype node from %d. Will follow reference" %
node)
# Choose reference with lowest id to avoid deletion
node = min(list(next_nodes.intersection(linear_ref_nodes)))
else:
# logging.warning("There is a deletion from node %d. Choosing lowest node id as next to avoid deletion." % node)
# This means more than one next node is on haplotype. Choose the one with lowest id to avoid taking deletion
node = min(list(next_on_haplotype))
logging.info("Found %d nodes. %d on haplotype" %
(len(nodes), n_haplotype_nodes))
haplotype_interval = Interval(0, graph.blocks[nodes[-1]].length(),
nodes, graph)
print("Path length: %d" % haplotype_interval.length())
file_base_name = out_base_name + "_" + str(haplotype)
IntervalCollection([haplotype_interval
]).to_file(file_base_name + ".intervalcollection",
text_file=True)
sequence = sequence_graph.get_interval_sequence(haplotype_interval)
out_file = open(file_base_name + ".fasta", "w")
out_file.writelines([">%s\n" % chrom])
out_file.writelines([sequence + "\n"])
out_file.close()
logging.info("Wrote fasta sequence to %s" % file_base_name + ".fasta")
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
def do_asserts(self):
for i, chromosome in enumerate(self.chromosomes):
final_peaks = IntervalCollection.create_list_from_file(
"multigraphs_" + chromosome + "_max_paths.intervalcollection")
for peak in self.peaks[i]:
assert peak in final_peaks
def _create_reads(self, *args):
super(TestMultipleGraphsCallPeaksCommandLine,
self)._create_reads(*args)
for intervals, chrom in zip(self.sample_reads, self.chromosomes):
IntervalCollection(intervals._intervals).to_file(
"test_sample_" + chrom + ".intervalcollection", text_file=True)
def predict_path(self):
logging.info("Using linear bonus %d on chromosome %s" % (self.linear_ref_bonus, self.chromosome))
logging.info("Using linear out base name %s" % self.out_file_base_name)
out_file = open("%s_%s.fasta" % (self.out_file_base_name, self.chromosome), "w")
# Traverse
first_nodes = self.graph.get_first_blocks()
assert len(first_nodes) == 1
logging.info("N nodes in graph: %d" % len(self.graph.blocks))
node = first_nodes[0]
assert node in self.linear_path_nodes, "Start node should be in linear ref"
path = []
n_ambigious = 0
edges_chosen = set()
i = 0
n_special_case = 0
while True:
if i % 1000000 == 0:
logging.info("%d nodes in graph traversed on chrom %s" % (i, self.chromosome))
i += 1
if self.max_nodes_to_traverse is not None and i > self.max_nodes_to_traverse:
logging.warning("Stopped traversing before end because max node to traverse was set")
break
path.append(node)
next_nodes = self.graph.adj_list[node]
if len(next_nodes) == 0:
logging.info("Done on node %d" % node)
break
elif len(next_nodes) == 1:
node = next_nodes[0]
else:
most_reads = 0
most_reads_node = next_nodes[0]
has_found_candidate_on_linear_ref = False
for next_node in next_nodes:
n_reads = self.edge_counts["%s-%s" % (node, next_node)]
if next_node in self.linear_path_nodes:
n_reads += self.linear_ref_bonus
if n_reads > most_reads or (n_reads >= most_reads and next_node in self.linear_path_nodes):
if node not in self.linear_path_nodes:
n_special_case += 1
# If already found something on linear ref, and this does not have more reads or lower id (not insertion), ignore
if has_found_candidate_on_linear_ref and n_reads == most_reads and next_node > most_reads_node:
continue # Ignore this alternative
most_reads_node = next_node
most_reads = n_reads
if next_node in self.linear_path_nodes:
has_found_candidate_on_linear_ref = True
if most_reads == 0:
n_ambigious += 1
assert most_reads_node is not None
edges_chosen.add("%d-%d" % (node, most_reads_node))
node = most_reads_node
if most_reads == 0:
# Assert we have taken linear ref path if exists
if any([n in self.linear_path_nodes for n in next_nodes]):
if node not in self.linear_path_nodes:
logging.error("Chose node %d as next, but it is not in linear ref." % node)
logging.error("Next nodes are: %s" % next_nodes)
for next_node in next_nodes:
if next_node in self.linear_path_nodes:
logging.error(" Node %d is in linear ref" % next_node)
else:
logging.error(" Node %d is not in linear ref" % next_node)
raise Exception("Could not traverse correctly")
# Find statistics of chosen nodes
nodes_chosen = set(path)
n_on_linear = len(nodes_chosen.intersection(self.linear_path_nodes))
n_not_on_linear = len(nodes_chosen) - n_on_linear
linear_ref_interval = Interval(0, self.graph.blocks[path[-1]].length(), path, self.graph)
IntervalCollection([linear_ref_interval]).to_file("%s_%s.intervalcollection" % (self.out_file_base_name, self.chromosome),
text_file=True)
logging.info("=== STATS FOR CHROMOSOME %s ===" % self.chromosome)
logging.info("N ambigious choices: %d" % n_ambigious)
logging.info("Total nodes in linear ref: %d" % len(self.linear_path_nodes))
logging.info("N nodes chosen that are not in linear ref: %d " % n_not_on_linear)
logging.info("N nodes chosen that are in linear ref: %d " % n_on_linear)
logging.info("N special case: %d" % n_special_case)
logging.info("N nodes in path: %d" % len(path))
logging.info("Linear path length: %d" % linear_ref_interval.length())
sequence = self.sequence_graph.get_interval_sequence(linear_ref_interval)
out_file.writelines([">%s\n" % self.chromosome])
out_file.writelines([sequence + "\n"])
out_file.close()
