def load_orthofinder_cluster(_run_id, ogfile):
"""Load cluster file from mcl (via orthofinder) into homology database"""
nseqs = 0
hist = {}
database.execute("BEGIN")
database.execute("DELETE FROM homology WHERE run_id=?;", (_run_id,))
with open(ogfile, 'r') as f:
cluster_id = 0
for line in f:
cluster = filter(lambda s : s.isdigit(), re.split(r'[@\s]', line))
n = len(cluster)
hist[n] = hist.get(n, 0) + 1
if n >= 4:
nseqs += n
for seq_id in cluster:
database.execute("""
INSERT INTO homology (run_id, component_id, sequence_id)
VALUES (?,?,?);""",
(_run_id, cluster_id, seq_id))
cluster_id += 1
database.execute("COMMIT")
utils.info(
"histogram of gene cluster sizes:\n",
'\n '.join("%d\t:\t%d" % (k, hist[k]) for k in sorted(hist)))
diagnostics.log('nseqs', nseqs)
diagnostics.log('histogram', hist)
def load_mcl_cluster(_run_id, cluster_file):
"""Load cluster file from mcl into homology database"""
nseqs = 0
hist = {}
database.execute("BEGIN")
database.execute("DELETE FROM homology WHERE run_id=?;", (_run_id,))
with open(cluster_file, "r") as f:
cluster_id = 0
for line in f:
cluster = filter(lambda s: s[0].isdigit(), line.rstrip().split())
n = len(cluster)
hist[n] = hist.get(n, 0) + 1
if n >= 4:
nseqs += n
for seq_id in cluster:
database.execute(
"""
INSERT INTO homology (run_id, component_id, sequence_id)
VALUES (?,?,?);""",
(_run_id, cluster_id, seq_id),
)
cluster_id += 1
database.execute("COMMIT")
utils.info("histogram of gene cluster sizes:\n", "\n ".join("%d\t:\t%d" % (k, hist[k]) for k in sorted(hist)))
diagnostics.log("nseqs", nseqs)
diagnostics.log("histogram", hist)
def component_histogram(self):
"""
Distribution of the number of nodes in each cluster.
"""
hist = {}
sql = """
SELECT component_id, COUNT(*)
FROM homology
WHERE run_id=?
GROUP BY component_id;"""
for _, count in database.execute(sql, (self.run_id,)):
hist[count] = hist.get(count, 0) + 1
if hist:
hist = numpy.array([(k, hist[k]) for k in sorted(hist.iterkeys())])
imgname = "%d.component.hist.png" % self.run_id
props = {"title": "Distribution of Cluster Sizes", "xlabel": "# Nodes in Cluster", "ylabel": "Frequency"}
return [self.histogram_overlay(imgname, [hist], props=props)]
