def _makeGraphFromM5(self, m5FN, qver_get_func, ice_opts):
"""Construct a graph from a BLASR M5 file."""
alignGraph = nx.Graph()
for r in blasr_against_ref(output_filename=m5FN,
is_FL=True,
sID_starts_with_c=False,
qver_get_func=qver_get_func,
ece_penalty=ice_opts.ece_penalty,
ece_min_len=ice_opts.ece_min_len):
if r.qID == r.cID:
continue # self hit, ignore
if r.ece_arr is not None:
logging.debug("adding edge {0},{1}".format(r.qID, r.cID))
alignGraph.add_edge(r.qID, r.cID)
return alignGraph
def _makeGraphFromM5(self, m5FN, qver_get_func, ice_opts):
"""Construct a graph from a BLASR M5 file."""
alignGraph = nx.Graph()
for r in blasr_against_ref(output_filename=m5FN,
is_FL=True,
sID_starts_with_c=False,
qver_get_func=qver_get_func,
ece_penalty=ice_opts.ece_penalty,
ece_min_len=ice_opts.ece_min_len):
if r.qID == r.cID:
continue # self hit, ignore
if r.ece_arr is not None:
logging.debug("adding edge {0},{1}".format(r.qID, r.cID))
alignGraph.add_edge(r.qID, r.cID)
return alignGraph
def build_uc_from_partial(input_fasta, ref_fasta, out_pickle,
sa_file=None, ccs_fofn=None,
done_filename=None, blasr_nproc=12, use_finer_qv=False):
"""
Given an input_fasta file of non-full-length (partial) reads and
(unpolished) consensus isoforms sequences in ref_fasta, align reads to
consensus isoforms using BLASR, and then build up a mapping between
consensus isoforms and reads (i.e., assign reads to isoforms).
Finally, save
{isoform_id: [read_ids],
nohit: set(no_hit_read_ids)}
to an output pickle file.
ccs_fofn --- If None, assume no quality value is available,
otherwise, use QV from ccs_fofn.
blasr_nproc --- equivalent to blasr -nproc, number of CPUs to use
"""
input_fasta = realpath(input_fasta)
m5_file = input_fasta + ".blasr"
out_pickle = realpath(out_pickle)
if sa_file is None:
if op.exists(input_fasta + ".sa"):
sa_file = input_fasta + ".sa"
cmd = "blasr {i} ".format(i=real_upath(input_fasta)) + \
"{r} -bestn 5 ".format(r=real_upath(ref_fasta)) + \
"-nproc {n} -m 5 ".format(n=blasr_nproc) + \
"-maxScore -1000 -minPctIdentity 85 " + \
"-out {o} ".format(o=real_upath(m5_file))
if sa_file is not None and op.exists(sa_file):
cmd += "-sa {sa}".format(sa=real_upath(sa_file))
logging.info("CMD: {cmd}".format(cmd=cmd))
_out, _code, _msg = backticks(cmd)
if _code != 0:
errMsg = "Command failed: {cmd}\n{e}".format(cmd=cmd, e=_msg)
logging.error(errMsg)
raise RuntimeError(errMsg)
if ccs_fofn is None:
logging.info("Loading probability from model (0.01,0.07,0.06)")
probqv = ProbFromModel(.01, .07, .06)
else:
start_t = time.time()
if use_finer_qv:
logging.info("Loading QVs from {i} + {f} took {s} secs".format(f=ccs_fofn, i=input_fasta,\
s=time.time()-start_t))
probqv = ProbFromQV(input_fofn=ccs_fofn, fasta_filename=input_fasta)
else:
input_fastq = input_fasta[:input_fasta.rfind('.')] + '.fastq'
logging.info("Converting {i} + {f} --> {fq}".format(i=input_fasta, f=ccs_fofn, fq=input_fastq))
ice_fa2fq(input_fasta, ccs_fofn, input_fastq)
logging.info("Loading QVs from {fq} took {s} secs".format(fq=input_fastq, s=time.time()-start_t))
probqv = ProbFromFastq(input_fastq)
logging.info("Calling blasr_against_ref ...")
hitItems = blasr_against_ref(output_filename=m5_file,
is_FL=False,
sID_starts_with_c=True,
qvmean_get_func=probqv.get_mean,
qver_get_func=probqv.get_smoothed,
ece_penalty=1,
ece_min_len=20,
same_strand_only=False,
max_missed_start=200,
max_missed_end=50)
partial_uc = {} # Maps each isoform (cluster) id to a list of reads
# which can map to the isoform
seen = set() # reads seen
logging.info("Building uc from BLASR hits.")
for h in hitItems:
if h.ece_arr is not None:
if h.cID not in partial_uc:
partial_uc[h.cID] = set()
partial_uc[h.cID].add(h.qID)
seen.add(h.qID)
for k in partial_uc:
partial_uc[k] = list(partial_uc[k])
allhits = set(r.name.split()[0] for r in FastaReader(input_fasta))
logging.info("Counting reads with no hit.")
nohit = allhits.difference(seen)
logging.info("Dumping uc to a pickle: {f}.".format(f=out_pickle))
with open(out_pickle, 'w') as f:
dump({'partial_uc': partial_uc, 'nohit': nohit}, f)
os.remove(m5_file)
done_filename = realpath(done_filename) if done_filename is not None \
else out_pickle + '.DONE'
logging.debug("Creating {f}.".format(f=done_filename))
touch(done_filename)
def build_uc_from_partial(input_fasta, ref_fasta, out_pickle,
sa_file=None, ccs_fofn=None,
done_filename=None, blasr_nproc=12, use_finer_qv=False):
"""
Given an input_fasta file of non-full-length (partial) reads and
(unpolished) consensus isoforms sequences in ref_fasta, align reads to
consensus isoforms using BLASR, and then build up a mapping between
consensus isoforms and reads (i.e., assign reads to isoforms).
Finally, save
{isoform_id: [read_ids],
nohit: set(no_hit_read_ids)}
to an output pickle file.
ccs_fofn --- If None, assume no quality value is available,
otherwise, use QV from ccs_fofn.
blasr_nproc --- equivalent to blasr -nproc, number of CPUs to use
"""
input_fasta = realpath(input_fasta)
m5_file = input_fasta + ".blasr"
out_pickle = realpath(out_pickle)
if sa_file is None:
if op.exists(input_fasta + ".sa"):
sa_file = input_fasta + ".sa"
cmd = "blasr {i} ".format(i=real_upath(input_fasta)) + \
"{r} -bestn 5 ".format(r=real_upath(ref_fasta)) + \
"-nproc {n} -m 5 ".format(n=blasr_nproc) + \
"-maxScore -1000 -minPctIdentity 85 " + \
"-out {o} ".format(o=real_upath(m5_file))
if sa_file is not None and op.exists(sa_file):
cmd += "-sa {sa}".format(sa=real_upath(sa_file))
logging.info("CMD: {cmd}".format(cmd=cmd))
_out, _code, _msg = backticks(cmd)
if _code != 0:
errMsg = "Command failed: {cmd}\n{e}".format(cmd=cmd, e=_msg)
logging.error(errMsg)
raise RuntimeError(errMsg)
if ccs_fofn is None:
logging.info("Loading probability from model (0.01,0.07,0.06)")
probqv = ProbFromModel(.01, .07, .06)
else:
start_t = time.time()
if use_finer_qv:
logging.info("Loading QVs from {i} + {f} took {s} secs".format(f=ccs_fofn, i=input_fasta,\
s=time.time()-start_t))
probqv = ProbFromQV(input_fofn=ccs_fofn, fasta_filename=input_fasta)
else:
input_fastq = input_fasta[:input_fasta.rfind('.')] + '.fastq'
logging.info("Converting {i} + {f} --> {fq}".format(i=input_fasta, f=ccs_fofn, fq=input_fastq))
ice_fa2fq(input_fasta, ccs_fofn, input_fastq)
logging.info("Loading QVs from {fq} took {s} secs".format(fq=input_fastq, s=time.time()-start_t))
probqv = ProbFromFastq(input_fastq)
logging.info("Calling blasr_against_ref ...")
hitItems = blasr_against_ref(output_filename=m5_file,
is_FL=False,
sID_starts_with_c=True,
qvmean_get_func=probqv.get_mean,
qver_get_func=probqv.get_smoothed,
ece_penalty=1,
ece_min_len=10,
same_strand_only=False)
partial_uc = {} # Maps each isoform (cluster) id to a list of reads
# which can map to the isoform
seen = set() # reads seen
logging.info("Building uc from BLASR hits.")
for h in hitItems:
if h.ece_arr is not None:
if h.cID not in partial_uc:
partial_uc[h.cID] = set()
partial_uc[h.cID].add(h.qID)
seen.add(h.qID)
for k in partial_uc:
partial_uc[k] = list(partial_uc[k])
allhits = set(r.name.split()[0] for r in FastaReader(input_fasta))
logging.info("Counting reads with no hit.")
nohit = allhits.difference(seen)
logging.info("Dumping uc to a pickle: {f}.".format(f=out_pickle))
with open(out_pickle, 'w') as f:
dump({'partial_uc': partial_uc, 'nohit': nohit}, f)
os.remove(m5_file)
done_filename = realpath(done_filename) if done_filename is not None \
else out_pickle + '.DONE'
logging.debug("Creating {f}.".format(f=done_filename))
touch(done_filename)
def build_uc_from_partial(input_fasta,
ref_fasta,
out_pickle,
sa_file=None,
ccs_fofn=None,
done_filename=None,
blasr_nproc=12):
"""Align consensus isoforms in ref_fasta and reads in input_fasta,
and save mappings between isoforms and reads to out_pickle.
ccs_fofn --- If None, assume no quality value is available,
otherwise, use QV from ccs_fofn.
blasr_nproc --- equivalent to blasr -nproc, number of CPUs to use
"""
input_fasta = realpath(input_fasta)
m5_file = input_fasta + ".blasr"
out_pickle = realpath(out_pickle)
if sa_file is None:
if op.exists(input_fasta + ".sa"):
sa_file = input_fasta + ".sa"
cmd = "blasr {i} ".format(i=input_fasta) + \
"{r} -bestn 5 ".format(r=ref_fasta) + \
"-nproc {n} -m 5 ".format(n=blasr_nproc) + \
"-maxScore -1000 -minPctIdentity 85 -out {o} ".format(o=m5_file)
if sa_file is not None and op.exists(sa_file):
cmd += "-sa {sa}".format(sa=sa_file)
logging.info("CMD: {cmd}".format(cmd=cmd))
_out, _code, _msg = backticks(cmd)
if _code != 0:
errMsg = "Command failed: {cmd}\n{e}".format(cmd=cmd, e=_msg)
logging.error(errMsg)
raise RuntimeError(errMsg)
if ccs_fofn is None:
logging.info("Loading probability from model")
probqv = ProbFromModel(.01, .07, .06)
else:
logging.info("Loading probability from QV in {f}".format(f=ccs_fofn))
probqv = ProbFromQV(input_fofn=ccs_fofn, fasta_filename=input_fasta)
logging.info("Calling blasr_against_ref ...")
hitItems = blasr_against_ref(output_filename=m5_file,
is_FL=False,
sID_starts_with_c=True,
qver_get_func=probqv.get_smoothed,
ece_penalty=1,
ece_min_len=10,
same_strand_only=False)
partial_uc = {} # Maps each isoform (cluster) id to a list of reads
# which can map to the isoform
seen = set() # reads seen
logging.info("Building uc from BLASR hits.")
for h in hitItems:
if h.ece_arr is not None:
if h.cID not in partial_uc:
partial_uc[h.cID] = []
partial_uc[h.cID].append(h.qID)
seen.add(h.qID)
allhits = set(r.name.split()[0] for r in FastaReader(input_fasta))
logging.info("Counting reads with no hit.")
nohit = allhits.difference(seen)
logging.info("Dumping uc to a pickle: {f}.".format(f=out_pickle))
with open(out_pickle, 'w') as f:
dump({'partial_uc': partial_uc, 'nohit': nohit}, f)
os.remove(m5_file)
done_filename = realpath(done_filename) if done_filename is not None \
else out_pickle + '.DONE'
logging.debug("Creating {f}.".format(f=done_filename))
touch(done_filename)
def build_uc_from_partial(
input_fasta, ref_fasta, out_pickle, sa_file=None, ccs_fofn=None, done_filename=None, blasr_nproc=12
):
"""Align consensus isoforms in ref_fasta and reads in input_fasta,
and save mappings between isoforms and reads to out_pickle.
ccs_fofn --- If None, assume no quality value is available,
otherwise, use QV from ccs_fofn.
blasr_nproc --- equivalent to blasr -nproc, number of CPUs to use
"""
input_fasta = realpath(input_fasta)
m5_file = input_fasta + ".blasr"
out_pickle = realpath(out_pickle)
if sa_file is None:
if op.exists(input_fasta + ".sa"):
sa_file = input_fasta + ".sa"
cmd = (
"blasr {i} ".format(i=input_fasta)
+ "{r} -bestn 5 ".format(r=ref_fasta)
+ "-nproc {n} -m 5 ".format(n=blasr_nproc)
+ "-maxScore -1000 -minPctIdentity 85 -out {o} ".format(o=m5_file)
)
if sa_file is not None and op.exists(sa_file):
cmd += "-sa {sa}".format(sa=sa_file)
logging.info("CMD: {cmd}".format(cmd=cmd))
_out, _code, _msg = backticks(cmd)
if _code != 0:
errMsg = "Command failed: {cmd}\n{e}".format(cmd=cmd, e=_msg)
logging.error(errMsg)
raise RuntimeError(errMsg)
if ccs_fofn is None:
logging.info("Loading probability from model")
probqv = ProbFromModel(0.01, 0.07, 0.06)
else:
logging.info("Loading probability from QV in {f}".format(f=ccs_fofn))
probqv = ProbFromQV(input_fofn=ccs_fofn, fasta_filename=input_fasta)
logging.info("Calling blasr_against_ref ...")
hitItems = blasr_against_ref(
output_filename=m5_file,
is_FL=False,
sID_starts_with_c=True,
qver_get_func=probqv.get_smoothed,
ece_penalty=1,
ece_min_len=10,
same_strand_only=False,
)
partial_uc = {} # Maps each isoform (cluster) id to a list of reads
# which can map to the isoform
seen = set() # reads seen
logging.info("Building uc from BLASR hits.")
for h in hitItems:
if h.ece_arr is not None:
if h.cID not in partial_uc:
partial_uc[h.cID] = []
partial_uc[h.cID].append(h.qID)
seen.add(h.qID)
allhits = set(r.name.split()[0] for r in FastaReader(input_fasta))
logging.info("Counting reads with no hit.")
nohit = allhits.difference(seen)
logging.info("Dumping uc to a pickle: {f}.".format(f=out_pickle))
with open(out_pickle, "w") as f:
dump({"partial_uc": partial_uc, "nohit": nohit}, f)
os.remove(m5_file)
done_filename = realpath(done_filename) if done_filename is not None else out_pickle + ".DONE"
logging.debug("Creating {f}.".format(f=done_filename))
touch(done_filename)
