def _download_mirbase(args, version="CURRENT"):
"""
Download files from mirbase
"""
if not args.hairpin or not args.mirna:
logger.info("Working with version %s" % version)
hairpin_fn = op.join(op.abspath(args.out), "hairpin.fa.gz")
mirna_fn = op.join(op.abspath(args.out), "miRNA.str.gz")
if not file_exists(hairpin_fn):
cmd_h = "wget ftp://mirbase.org/pub/mirbase/%s/hairpin.fa.gz -O %s && gunzip -f !$" % (
version, hairpin_fn)
do.run(cmd_h, "download hairpin")
if not file_exists(mirna_fn):
cmd_m = "wget ftp://mirbase.org/pub/mirbase/%s/miRNA.str.gz -O %s && gunzip -f !$" % (
version, mirna_fn)
do.run(cmd_m, "download mirna")
else:
return args.hairpin, args.mirna
def miraligner(args):
"""
Realign BAM hits to miRBAse to get better accuracy and annotation
"""
hairpin, mirna = _download_mirbase(args)
precursors = _read_precursor(args.hairpin, args.sps)
matures = _read_mature(args.mirna, args.sps)
gtf = _read_gtf(args.gtf)
out_dts = []
for bam_fn in args.files:
sample = op.splitext(op.basename(bam_fn))[0]
if bam_fn.endswith("bam") or bam_fn.endswith("sam"):
logger.info("Reading %s" % bam_fn)
bam_fn = _sam_to_bam(bam_fn)
bam_sort_by_n = _bam_sort(bam_fn)
reads = _read_bam(bam_sort_by_n, precursors)
elif bam_fn.endswith("fasta") or bam_fn.endswith(
"fa") or bam_fn.endswith("fastq"):
out_file = op.join(args.out, sample + ".premirna")
bam_fn = _filter_seqs(bam_fn)
if args.miraligner:
_cmd_miraligner(bam_fn, out_file, args.sps, args.hairpin)
reads = _read_miraligner(out_file)
else:
logger.error("No other aligner supported")
raise ValueError("Specify --miraligner or give BAM files")
else:
raise ValueError("Format not recognized.")
if not args.miraligner:
reads = _annotate(reads, matures, precursors)
out_file = op.join(args.out, sample + ".mirna")
out_file, dt, dt_pre = _tab_output(reads, out_file, sample)
try:
vcf_file = op.join(args.out, sample + ".vcf")
if not file_exists(vcf_file):
# if True:
create_vcf(dt_pre, matures, gtf, vcf_file)
try:
import vcf
vcf.Reader(filename=vcf_file)
except Exception as e:
logger.warning(e.__doc__)
logger.warning(e.message)
except Exception as e:
# traceback.print_exc()
logger.warning(e.__doc__)
logger.warning(e.message)
if isinstance(dt, pd.DataFrame):
out_dts.append(dt)
if out_dts:
_create_counts(out_dts, args.out)
# _summarize(out_dts)
else:
print "No files analyzed!"
def _cmd_miraligner(fn, out_file, species, hairpin):
"""
Run miraligner for miRNA annotation
"""
tool = _get_miraligner()
path_db = op.dirname(op.abspath(hairpin))
opts = "-Xms750m -Xmx4g"
cmd = "{tool} -freq -i {fn} -o {out_file} -s {species} -db {path_db} -sub 1 -trim 3 -add 3"
if not file_exists(out_file):
do.run(cmd.format(**locals()), "miraligner with %s" % fn)
shutil.move(out_file + ".mirna", out_file)
return out_file
def _filter_seqs(fn):
"""Convert names of sequences to unique ids"""
out_file = op.splitext(fn)[0] + "_unique.fa"
idx = 0
if not file_exists(out_file):
with open(out_file, 'w') as out_handle:
with open(fn) as in_handle:
for line in in_handle:
if line.startswith("@") or line.startswith(">"):
fixed_name = _make_unique(line.strip(), idx)
seq = in_handle.next().strip()
counts = _get_freq(fixed_name)
if len(seq) < 26 and (counts > 1 or counts == 0):
idx += 1
print >> out_handle, fixed_name
print >> out_handle, seq
try:
if line.startswith("@"):
in_handle.next()
in_handle.next()
except:
pass
return out_file
def _bam_sort(bam_fn):
bam_sort_by_n = op.splitext(bam_fn)[0] + "_sort.bam"
if not file_exists(bam_sort_by_n):
do.run(("samtools sort -n -o {bam_sort_by_n} {bam_fn}").format(
**locals()))
return bam_sort_by_n
def _bam_sort(bam_fn):
bam_sort_by_n = op.splitext(bam_fn)[0] + "_sort.bam"
if not file_exists(bam_sort_by_n):
do.run(("samtools sort -n -o {bam_sort_by_n} {bam_fn}").format(
**locals()))
return bam_sort_by_n