def process_self_align_into_seed(align_filename,
seqids,
reader_class,
pCS=None,
dun_use_partial=False):
"""
Ignore hits that are - strand or qID >= sID (self hit or already reported)
Returns:
pCS -- preClusterSet
orphans --- seqs that are neither in pCS nor tucked i.e. no align hits
"""
if pCS is None:
pCS = preClusterSet2()
orphans = set(seqids)
reader = reader_class(align_filename)
for r in reader:
if r.qID >= r.sID or r.strand == '-': continue
s = r.characterize(30, 0.01, 30, 0.01, 30, 0.05, min_identity=0.99)
if dun_use_partial and s == 'partial': continue
if s == 'match':
pCS.add_seqid_match(r.qID, r.sID)
if r.qID in bug_ids or r.sID in bug_ids:
print "match:", r.qID, r.sID
print "after match:", pCS.seq_stat[r.qID], pCS.seq_stat[r.sID]
elif s == 'partial':
pCS.add_seqid_partial(r.qID, r.sID)
if r.qID in bug_ids or r.sID in bug_ids:
print "partial:", r.qID, r.sID
print "after partial:", pCS.seq_stat[r.qID], pCS.seq_stat[
r.sID]
elif s == 'q_contained':
if r.qID in bug_ids or r.sID in bug_ids:
print "tucking {0} into {1}".format(r.qID, r.sID)
print "before:", pCS.seq_stat[r.qID], pCS.seq_stat[r.sID]
pCS.add_seqid_contained(r.qID, r.sID)
if r.qID in bug_ids or r.sID in bug_ids:
print "after:", pCS.seq_stat[r.qID], pCS.seq_stat[r.sID]
elif s == 's_contained':
if r.qID in bug_ids or r.sID in bug_ids:
print "tucking {0} into {1}".format(r.sID, r.qID)
print "before:", pCS.seq_stat[r.sID], pCS.seq_stat[r.qID]
pCS.add_seqid_contained(r.sID, r.qID)
if r.qID in bug_ids or r.sID in bug_ids:
print "after:", pCS.seq_stat[r.sID], pCS.seq_stat[r.qID]
try:
orphans.remove(r.qID)
except:
pass
try:
orphans.remove(r.sID)
except:
pass
#sanity_checking(pCS, orphans)
#sanity_checking(pCS, orphans)
return pCS, orphans
def process_self_align_into_seed(align_filename,
seqids,
reader_class,
pCS=None,
dun_use_partial=False):
"""
Ignore hits that are - strand or qID >= sID (self hit or already reported)
Returns:
pCS -- preClusterSet
orphans --- seqs that are neither in pCS nor tucked i.e. no align hits
"""
if pCS is None:
pCS = preClusterSet2()
orphans = set(seqids)
reader = reader_class(align_filename)
for r in reader:
if r.qID >= r.sID or r.strand == '-': continue
s = r.characterize(100, 0.05, 50, 0.02, 20, 0.01)
if dun_use_partial and s == 'partial': continue
#if r.qID=='m54119_170322_155415/12845906/28_5778_CCS' or r.sID=='m54119_170322_155415/12845906/28_5778_CCS':
# pdb.set_trace()
# Liz note: currently, just add all to match because minimap sensitivity not enough to do "tuck" properly
if s == 'match' or s == 'partial' or s.endswith('_contained'):
pCS.add_seqid_match(r.qID, r.sID)
#elif s == 'q_contained':
# pCS.add_seqid_contained(r.qID, r.sID)
#elif s == 's_contained':
# pCS.add_seqid_contained(r.sID, r.qID)
try:
orphans.remove(r.qID)
except:
pass
try:
orphans.remove(r.sID)
except:
pass
#sanity_checking(pCS, orphans)
#sanity_checking(pCS, orphans)
return pCS, orphans
def read_seq_csv(csv_filename):
# sanity check that "seqid" and "stat" are two valid column headers
header_checked = False
orphans = set()
pCS = preClusterSet2()
for r in DictReader(open(csv_filename), delimiter=','):
if not header_checked:
if 'seqid' not in r or 'stat' not in r:
print("{0} must have the fields 'seqid' and 'stat'! Abort".format(csv_filename), file=sys.stderr)
sys.exit(-1)
header_checked = True
if r['stat']=='orphan':
orphans.add(r['seqid'])
else:
cid = int(r['stat'])
if cid not in pCS.S: pCS.S[cid] = preCluster(cid=cid)
pCS.add_seqid_to_cluster_by_cid(r['seqid'], cid)
return pCS, orphans
def process_self_align_into_seed(align_filename,
seqids,
reader_class,
pCS=None):
"""
Ignore hits that are - strand or qID >= sID (self hit or already reported)
Returns:
pCS -- preClusterSet
orphans --- seqs that are neither in pCS nor tucked i.e. no align hits
"""
if pCS is None:
pCS = preClusterSet2()
orphans = set(seqids)
reader = reader_class(align_filename)
for r in reader:
if r.qID >= r.sID or r.strand == '-': continue
s = r.characterize(400, 0.1, 400, 0.1, 100, 0.1)
if s == 'partial': continue
#if r.qID=='m54119_170322_155415/12845906/28_5778_CCS' or r.sID=='m54119_170322_155415/12845906/28_5778_CCS':
# pdb.set_trace()
if s == 'match':
pCS.add_seqid_match(r.qID, r.sID)
elif s == 'q_contained':
pCS.add_seqid_contained(r.qID, r.sID)
elif s == 's_contained':
pCS.add_seqid_contained(r.sID, r.qID)
try:
orphans.remove(r.qID)
except:
pass
try:
orphans.remove(r.sID)
except:
pass
#sanity_checking(pCS, orphans)
#sanity_checking(pCS, orphans)
return pCS, orphans
def process_self_align_into_seed(align_filename, seqids, reader_class, pCS=None, dun_use_partial=False):
"""
Ignore hits that are - strand or qID >= sID (self hit or already reported)
Returns:
pCS -- preClusterSet
orphans --- seqs that are neither in pCS nor tucked i.e. no align hits
"""
if pCS is None:
pCS = preClusterSet2()
orphans = set(seqids)
reader = reader_class(align_filename)
for r in reader:
if r.qID >= r.sID or r.strand == '-': continue
s = r.characterize(400, 0.4, 100, 0.1, 50, 0.05)
if dun_use_partial and s == 'partial': continue
#if r.qID=='m54119_170322_155415/12845906/28_5778_CCS' or r.sID=='m54119_170322_155415/12845906/28_5778_CCS':
# pdb.set_trace()
# Liz note: currently, just add all to match because minimap sensitivity not enough to do "tuck" properly
if s == 'match' or s == 'partial' or s.endswith('_contained'):
pCS.add_seqid_match(r.qID, r.sID)
#elif s == 'q_contained':
# pCS.add_seqid_contained(r.qID, r.sID)
#elif s == 's_contained':
# pCS.add_seqid_contained(r.sID, r.qID)
try:
orphans.remove(r.qID)
except:
pass
try:
orphans.remove(r.sID)
except:
pass
#sanity_checking(pCS, orphans)
#sanity_checking(pCS, orphans)
return pCS, orphans