def _check_line(line, num, num_samples):
""" Check file for minimum
"""
fields = read_gff_line(line)
# Check seqID
if not fields['chrom']:
logger.error('MISSING seqID in line %s' % (num))
# Check source
source = (fields['source']).lower()
valid_source = False
valid_sources = ["mirBase", "mirgeneDB"]
if (any(s.lower() in source for s in valid_sources)):
valid_source = True
if valid_source is False:
logger.error('INCORRECT SOURCE in line %s' % (num))
# Check type
type = fields['type']
source = (fields['source']).lower()
valid_type = False
if type in ["ref_miRNA", "isomiR"]:
valid_type = True
if valid_type is False:
logger.error('INCORRECT TYPE in line %s' % (num))
# Check start/end
if not fields['start']:
logger.error('MISSING START value in line %s' % (num))
if not fields['end']:
logger.error('MISSING END value in line %s' % (num))
# Check strand
if str(fields['strand']) not in ["+", "-"]:
logger.error('INCORRECT STRAND in line %s' % (num))
# Check attribute-variant
variant = (fields['attrb']['Variant']).lower()
valid_variant = False
valid_variants = version.GFFv[version.current]
if (any(s.lower() in variant for s in valid_variants)):
valid_variant = True
if valid_variant is False:
logger.error('INCORRECT VARIANT type in line %s' % (num))
# Check attribute-expression
expression = fields['attrb']['Expression'].strip().split(",")
expression = filter(None, expression)
if len(expression) != num_samples:
logger.error('INCORRECT number of EXPRESSION VALUES \
in line %s' % (num))
def _check_line(line, num, num_samples):
""" Check file for minimum
"""
fields = read_gff_line(line)
# Check seqID
if not fields['chrom']:
logger.error('MISSING seqID in line %s' % (num))
# Check source
source = (fields['source']).lower()
valid_source = False
valid_sources = ["mirBase", "mirgeneDB"]
if (any(s.lower() in source for s in valid_sources)):
valid_source = True
if valid_source is False:
logger.error('INCORRECT SOURCE in line %s' % (num))
# Check type
type = fields['type']
source = (fields['source']).lower()
valid_type = False
if type in ["ref_miRNA", "isomiR"]:
valid_type = True
if valid_type is False:
logger.error('INCORRECT TYPE in line %s' % (num))
# Check start/end
if not fields['start']:
logger.error('MISSING START value in line %s' % (num))
if not fields['end']:
logger.error('MISSING END value in line %s' % (num))
# Check strand
if str(fields['strand']) not in ["+", "-"]:
logger.error('INCORRECT STRAND in line %s' % (num))
# Check attribute-variant
variant = (fields['attrb']['Variant']).lower()
valid_variant = False
valid_variants = version.GFFv[version.current]
if (any(s.lower() in variant for s in valid_variants)):
valid_variant = True
if valid_variant is False:
logger.error('INCORRECT VARIANT type in line %s' % (num))
# Check attribute-expression
expression = fields['attrb']['Expression'].strip().split(",")
expression = filter(None, expression)
if len(expression) != num_samples:
logger.error('INCORRECT number of EXPRESSION VALUES \
in line %s' % (num))
def read_reference(fn):
"""Read GFF into UID:Variant
Args:
*fn (str)*: GFF file.
Returns:
*srna (dict)*: dict with >>> {'UID': 'iso_snp:-2,...'}
"""
srna = dict()
with open(fn) as inh:
for line in inh:
if line.startswith("#"):
continue
cols = read_gff_line(line)
attr = cols['attrb']
srna[attr['UID']] = [_simplify(attr['Variant']), attr]
return srna
def read_reference(fn):
"""Read GFF into UID:Variant
Args:
*fn (str)*: GFF file.
Returns:
*srna (dict)*: dict with >>> {'UID': 'iso_snp:-2,...'}
"""
srna = dict()
with open(fn) as inh:
for line in inh:
if line.startswith("#"):
continue
cols = read_gff_line(line)
attr = cols['attrb']
srna[attr['UID']] = [_simplify(attr['Variant']), attr]
return srna
def _compare_to_reference(fn, reference):
same = 0
diff = list()
extra = list()
miss = list()
results = list()
seen = 0
seen_reference = set()
with open(fn) as inh:
for line in inh:
if line.startswith("#"):
continue
cols = read_gff_line(line)
attr = cols['attrb']
if attr['UID'] in reference:
mirna = "Y" if attr['Name'] == reference[
attr['UID']][1]['Name'] else attr['Name']
accuracy = _accuracy(_simplify(attr['Variant']),
reference[attr['UID']][0])
results.append([attr['UID'], "D", mirna, accuracy])
if _simplify(attr['Variant']) == reference[attr['UID']][0]:
same += 1
else:
diff.append("%s | reference: %s" %
(line.strip(), reference[attr['UID']][1]))
seen += 1
seen_reference.add(attr['UID'])
else:
extra.append("%s | extra" % line.strip())
results.append([
attr['UID'], "E", attr['Name'],
_accuracy(_simplify(attr['Variant']), "")
])
for uid in reference:
if uid not in seen_reference:
results.append([uid, "M", "N", _accuracy("", reference[uid][0])])
miss.append("| miss %s" % reference[uid][1])
logger.info("Number of sequences found in reference: %s" % seen)
logger.info("Number of sequences matches reference: %s" % same)
logger.info("Number of sequences different than reference: %s" % len(diff))
logger.info("Number of sequences extra sequences: %s" % len(extra))
logger.info("Number of sequences missed sequences: %s" % len(miss))
return results
def _calc_stats(fn):
"""
Read files and parse into categories
"""
samples = _get_samples(fn)
lines = []
seen = set()
with open(fn) as inh:
for line in inh:
if line.startswith("#"):
continue
cols = read_gff_line(line)
logger.debug("## STATS: attribute %s" % cols['attrb'])
attr = cols['attrb']
if attr['Filter'] != "Pass":
continue
if "-".join([attr['UID'], attr['Variant'], attr['Name']]) in seen:
continue
seen.add("-".join([attr['UID'], attr['Variant'], attr['Name']]))
lines.extend(_classify(cols['type'], attr, samples))
df = _summary(lines)
return df
def _compare_to_reference(fn, reference):
same = 0
diff = list()
extra = list()
miss = list()
results = list()
seen = 0
seen_reference = set()
with open(fn) as inh:
for line in inh:
if line.startswith("#"):
continue
cols = read_gff_line(line)
attr = cols['attrb']
if attr['UID'] in reference:
mirna = "Y" if attr['Name'] == reference[attr['UID']][1]['Name'] else attr['Name']
accuracy = _accuracy(_simplify(attr['Variant']), reference[attr['UID']][0])
results.append([attr['UID'], "D", mirna, accuracy])
if _simplify(attr['Variant']) == reference[attr['UID']][0]:
same += 1
else:
diff.append("%s | reference: %s" % (line.strip(), reference[attr['UID']][1]))
seen += 1
seen_reference.add(attr['UID'])
else:
extra.append("%s | extra" % line.strip())
results.append([attr['UID'], "E", attr['Name'], _accuracy(_simplify(attr['Variant']), "")])
for uid in reference:
if uid not in seen_reference:
results.append([uid, "M", "N", _accuracy("", reference[uid][0])])
miss.append("| miss %s" % reference[uid][1])
logger.info("Number of sequences found in reference: %s" % seen)
logger.info("Number of sequences matches reference: %s" % same)
logger.info("Number of sequences different than reference: %s" % len(diff))
logger.info("Number of sequences extra sequences: %s" % len(extra))
logger.info("Number of sequences missed sequences: %s" % len(miss))
return results
def _fix(line, expression):
# Need to fix Read attribute since not usefull when multiple sample in a line.
cols = read_gff_line(line)
cols['attrb']['Expression'] = expression
return paste_columns(cols, guess_format(line))
def test_read_line(self):
"""Read GFF/GTF line"""
from mirtop.gff.body import read_gff_line
with open("data/examples/gff/2samples.gff") as inh:
for line in inh:
print(read_gff_line(line))
def convert_gff_counts(args):
""" Reads a GFF file to produces output file containing Expression counts
Args:
*args(namedtuple)*: arguments parsed from command line with
*mirtop.libs.parse.add_subparser_counts()*.
Returns:
*file (file)*: with columns like:
UID miRNA Variant Sample1 Sample2 ... Sample N
"""
sep = "\t"
variant_header = sep.join(['iso_5p', 'iso_3p',
'iso_add', 'iso_snp'])
if args.add_extra:
precursors = fasta.read_precursor(args.hairpin, args.sps)
matures = mapper.read_gtf_to_precursor(args.gtf)
variant_header = sep.join([variant_header,
'iso_5p_nt', 'iso_3p_nt',
'iso_add_nt', 'iso_snp_nt'])
logger.info("INFO Reading GFF file %s", args.gff)
logger.info("INFO Writing TSV file to directory %s", args.out)
gff_file = open(args.gff, 'r')
out_file = op.join(args.out, "expression_counts.tsv")
missing_parent = 0
missing_mirna = 0
unvalid_uid = 0
with open(out_file, 'w') as outh:
for samples_line in gff_file:
if samples_line.startswith("## COLDATA:"):
samples = sep.join(samples_line.strip().split("COLDATA:")[1].strip().split(","))
header = sep.join(['UID', 'Read', 'miRNA', 'Variant',
variant_header, samples])
print(header, file=outh)
break
for mirna_line in gff_file:
mirna_values = read_gff_line(mirna_line)
Read = mirna_values["attrb"]["Read"]
UID = mirna_values["attrb"]["UID"]
mirna = mirna_values["attrb"]["Name"]
parent = mirna_values["attrb"]["Parent"]
variant = mirna_values["attrb"]["Variant"]
try:
read_id(UID)
except KeyError:
unvalid_uid += 1
continue
expression = sep.join(mirna_values["attrb"]["Expression"].strip().split(","))
cols_variants = sep.join(_expand(variant))
logger.debug("COUNTS::Read:%s" % Read)
logger.debug("COUNTS::EXTRA:%s" % variant)
if args.add_extra:
if parent not in precursors:
missing_parent += 1
continue
if mirna not in matures[parent]:
missing_mirna += 1
continue
extra = variant_with_nt(mirna_line, precursors, matures)
if extra == "Invalid":
continue
logger.debug("COUNTS::EXTRA:%s" % extra)
cols_variants = sep.join([cols_variants] + _expand(extra, True))
summary = sep.join([UID, Read, mirna, variant,
cols_variants, expression])
logger.debug(summary)
print(summary, file=outh)
gff_file.close()
logger.info("Missing Parents in hairpin file: %s" % missing_parent)
logger.info("Missing MiRNAs in GFF file: %s" % missing_mirna)
logger.info("Non valid UID: %s" % unvalid_uid)
logger.info("Output file is at %s" % out_file)
def _fix(line, expression):
# Need to fix Read attribute since not usefull when multiple sample in a line.
cols = read_gff_line(line)
cols['attrb']['Expression'] = expression
return paste_columns(cols, guess_format(line))
def read_file(fn, args):
"""
Read isomiR-SEA file and convert to mirtop GFF format.
Args:
*fn(str)*: file name with isomiR-SEA output information.
*database(str)*: database name.
*args(namedtuple)*: arguments from command line.
See *mirtop.libs.parse.add_subparser_gff()*.
Returns:
*reads (nested dicts)*:gff_list has the format as
defined in *mirtop.gff.body.read()*.
"""
database = args.database
gtf = args.gtf
sep = " " if args.out_format == "gtf" else "="
map_mir = mapper.read_gtf_to_mirna(gtf)
reads = defaultdict(dict)
reads_in = 0
sample = os.path.splitext(os.path.basename(fn))[0]
hits = _get_hits(fn)
logger.debug("ISOMIRSEA::SAMPLE::%s" % sample)
with open(fn) as handle:
for line in handle:
cols = line.strip().split("\t")
attr = read_attributes(line, "=")
query_name = attr['TS']
query_sequence = attr['TS'].replace("U", "T")
start = int(cols[3])
end = int(cols[4])
isomirseq_iso = attr['ISO']
if query_name not in reads and query_sequence == None:
continue
if query_sequence and query_sequence.find("N") > -1:
continue
counts = attr["TC"]
chrom = cols[0]
# logger.debug("SEQBUSTER:: cigar {cigar}".format(**locals()))
cigar = attr['CI'].replace("U", "T")
idu = make_id(query_sequence)
isoformat = cigar2variants(cigar, query_sequence, attr['ISO'])
logger.debug("\nISOMIRSEA::NEW::query: {query_sequence}\n"
" precursor {chrom}\n"
" name: {query_name}\n"
" idu: {idu}\n"
" start: {start}\n"
" cigar: {cigar}\n"
" iso: {isoformat}\n"
" variant: {isoformat}".format(**locals()))
source = "isomiR" if isoformat != "NA" else "ref_miRNA"
strand = "+"
database = cols[1]
mirName = attr['MIN'].split()[0]
preName = attr['PIN'].split()[0]
score = "."
Filter = attr['FILTER']
isotag = attr['ISO']
tchrom, tstart = _genomic2transcript(map_mir[mirName],
chrom, start)
start = start if not tstart else tstart
chrom = chrom if not tstart else tchrom
end = start + len(query_sequence)
hit = hits[idu]
attrb = ("Read {query_sequence}; UID {idu}; Name {mirName};"
" Parent {preName}; Variant {isoformat};"
" Isocode {isotag}; Cigar {cigar}; Expression {counts};"
" Filter {Filter}; Hits {hit};").format(**locals())
line = ("{chrom}\t{database}\t{source}\t{start}\t{end}\t"
"{score}\t{strand}\t.\t{attrb}").format(**locals())
if args.add_extra:
extra = variant_with_nt(line, args.precursors, args.matures)
line = "%s Changes %s;" % (line, extra)
line = paste_columns(read_gff_line(line), sep=sep)
if start not in reads[chrom]:
reads[chrom][start] = []
if Filter == "Pass":
reads_in += 1
reads[chrom][start].append([idu, chrom, counts, sample, line])
logger.info("Hits: %s" % reads_in)
return reads
def test_read_line(self):
"""Read GFF/GTF line"""
from mirtop.gff.body import read_gff_line
with open("data/examples/gff/2samples.gff") as inh:
for line in inh:
print(read_gff_line(line))
def read_file(folder, args):
"""
Read sRNAbench file and convert to mirtop GFF format.
Args:
*fn(str)*: file name with sRNAbench output information.
*database(str)*: database name.
*args(namedtuple)*: arguments from command line.
See *mirtop.libs.parse.add_subparser_gff()*.
Returns:
*reads (nested dicts)*:gff_list has the format as
defined in *mirtop.gff.body.read()*.
"""
reads_anno = os.path.join(folder, "reads.annotation")
reads_iso = os.path.join(folder, "microRNAannotation.txt")
sep = " " if args.out_format == "gtf" else "="
sample = os.path.basename(folder)
database = args.database
precursors = args.precursors
matures = args.matures
n_out = 0
n_in = 0
n_ns = 0
n_notassign = 0
n_notindb = 0
reads = defaultdict(dict)
seen = set()
source_iso = _read_iso(reads_iso)
logger.info("Reads with isomiR information %s" % len(source_iso))
with open(reads_anno) as handle:
for sequence in handle:
cols = sequence.strip().split("\t")
query_name = cols[0]
query_sequence = cols[0]
if query_name not in reads and not query_sequence:
continue
if query_sequence and query_sequence.find("N") > -1:
n_ns += 1
continue
if cols[3].find("mature") == -1:
n_in += 1
continue
counts = int(cols[1])
hit = len(set([mirna.split("#")[1] for mirna in cols[4].split("$")]))
for nhit in cols[4].split("$"):
logger.debug("SRNABENCH::line hit: %s" % nhit)
hit_info = nhit.split("#")
pos_info = hit_info[3].split(",")
start = int(pos_info[1]) - 1
end = start + len(query_sequence) # int(pos_info[2]) - 1
chrom = pos_info[0]
mirName = hit_info[1]
if chrom not in precursors or chrom not in matures:
n_notindb += 1
if mirName not in matures[chrom]:
n_notindb += 1
if (query_sequence, mirName) in seen:
continue
seen.add((query_sequence, mirName))
if (query_sequence, mirName) not in source_iso:
continue
isoformat = source_iso[(query_sequence, mirName)]
if isoformat == "mv":
n_notassign += 1
continue
source = "isomiR" if isoformat != "NA" else "ref_miRNA"
logger.debug("SRNABENCH::query: {query_sequence}\n"
" precursor {chrom}\n"
" name: {query_name}\n"
" start: {start}\n"
" external: {isoformat}\n"
" hit: {hit}".format(**locals()))
logger.debug("SRNABENCH:: start %s end %s" % (start, end))
if len(precursors[chrom]) < start + len(query_sequence):
n_out += 1
continue
Filter = "Pass"
cigar = make_cigar(query_sequence,
precursors[chrom][start:end])
preName = chrom
score = "."
strand = "+"
idu = make_id(query_sequence)
attrb = ("Read {query_sequence}; UID {idu}; Name {mirName};"
" Parent {preName}; Variant {isoformat};"
" Cigar {cigar}; Expression {counts};"
" Filter {Filter}; Hits {hit};").format(**locals())
line = ("{chrom}\t{database}\t{source}\t{start}\t{end}\t"
"{score}\t{strand}\t.\t{attrb}").format(**locals())
if args.add_extra:
extra = variant_with_nt(line, args.precursors,
args.matures)
line = "%s Changes %s;" % (line, extra)
line = paste_columns(read_gff_line(line), sep=sep)
if start not in reads[chrom]:
reads[chrom][start] = []
if Filter == "Pass":
n_in += 1
reads[chrom][start].append([idu, chrom, counts,
sample, line])
logger.info("Loaded %s reads with %s hits" % (len(reads), n_in))
logger.info("Reads without precursor information: %s" % n_notindb)
logger.info("Reads with MV as variant definition,"
" not supported by GFF: %s" % n_notassign)
logger.info("Hit Filtered by having > 3 changes: %s" % n_out)
return reads
def create_vcf(mirgff3, precursor, gtf, vcffile):
"""
Args:
'mirgff3(str)': File with mirGFF3 format that will be converted
'precursor(str)': Fasta format sequences of all miRNA hairpins
'gtf(str)': Genome coordinates
'vcffile': name of the file to be saved
Returns:
Nothing is returned, instead, a VCF file is generated
"""
#Check if the input files exist:
try:
gff3_file = open(mirgff3, "r", encoding="utf-8") if six.PY3 else open(
mirgff3, "r")
except IOError:
print("Can't read the file", end=mirgff3)
sys.exit()
with gff3_file:
data = gff3_file.read()
if six.PY2:
data = data.decode("utf-8-sig").encode("utf-8")
gff3_data = data.split("\n")
vcf_file = open(vcffile, "w")
ver = "v4.3" # Current VCF version formatting
vcf_file.write("##fileformat=VCF%s\n" % ver)
date = datetime.datetime.now().strftime("%Y%m%d")
vcf_file.write("##fileDate=%s\n" % date)
source = "\n".join(s for s in gff3_data
if "## source-ontology: " in s)[20:]
line = 0
sample_names = []
while gff3_data[line][:2] == "##":
if gff3_data[line][:19] == "## source-ontology:":
source = gff3_data[line][20:]
elif gff3_data[line][:11] == "## COLDATA:":
sample_names = gff3_data[line][12:].split(",")
line += 1
vcf_file.write("##source=%s\n" % source)
vcf_file.write(
'##INFO=<ID=NS,Type=Integer,Description="Number of samples"\n')
vcf_file.write("##FILTER=<ID=REJECT,Description='"
'Filter not passed'
"'>\n")
vcf_file.write(
'##FORMAT=<ID=TRC,Number=1,Type=Integer,Description="Total read count">\n'
)
vcf_file.write(
'##FORMAT=<ID=TSC,Number=1,Type=Integer,Description="Total SNP count">\n'
)
vcf_file.write(
'##FORMAT=<ID=TMC,Number=1,Type=Integer,Description="Total miRNA count">\n'
)
vcf_file.write(
'##FORMAT=<ID=GT,Number=1,Type=Integer,Description="Genotype">\n')
header = "#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT"
# Adds Header
for s in range(len(sample_names)):
header = header + "\t" + sample_names[s]
vcf_file.write(header)
all_dict = dict(
) # initializing an empty dictionary where all info will be added
key_list = [
] # Initializing a list which will contain all the keys of the dictionary
mirna_dict = dict(
) # initializing an empty dictionary where mirna info will be put
n_SNP = 0
n_noSNP = 0
no_var = 0
hairpins = read_precursor(precursor)
gff3 = read_gtf_to_precursor(gtf)
gtf_dic = read_gtf_to_mirna(gtf)
for line in range(0, len(gff3_data)):
if not gff3_data[line]:
continue
if gff3_data[line][1] == "#":
continue
else: # Parsing the gff3 mirna lecture:
gff_fields = read_gff_line(gff3_data[line])
gtf_name = gff_fields['attrb']['Name']
gtf_parent = gff_fields['attrb']['Parent']
if gtf_parent not in gff3:
continue
if gtf_name not in gff3[gtf_parent]:
continue
parent_ini_pos = gff3[gtf_parent][gtf_name][0]
parent_end_pos = gff3[gtf_parent][gtf_name][1]
ref_seq = (hairpins[gtf_parent][parent_ini_pos:parent_end_pos + 1])
vcf_chrom = gtf_dic[gtf_name][gtf_parent][0]
vcf_pos = int(gff_fields['start']) + int(
gtf_dic[gtf_name][gtf_parent][1])
hairpin = hairpins[gtf_parent]
variants = gff_fields['attrb']['Variant'].split(",")
logger.debug("VCF::Variant::%s" % variants)
# Obtaining the iso_3p, iso_add3p and iso_5p values:
var3p = [s for s in variants if 'iso_3p' in s]
if len(var3p):
var3p = int(var3p[0][7:]) # Position of iso_3p value
else:
var3p = 0
var_add3p = [s for s in variants if 'iso_add3p' in s]
if len(var_add3p):
var_add3p = int(
var_add3p[0][10:]) # Position of iso_add3p value
else:
var_add3p = 0
var3p = var3p + var_add3p
logger.debug("VCF::VAR_3p::%s" % var3p)
var5p = [s for s in variants if 'iso_5p' in s]
if len(var5p):
var5p = int(var5p[0][7:]) # Position of iso_5p value
else:
var5p = 0 #
logger.debug("VCF::VAR_5p::%s" % var5p)
cigar = gff_fields['attrb']["Cigar"]
# Obtaining all the variants from the cigar:
if 1:
(key_pos, key_var, vcf_ref, vcf_alt) = cigar_2_key(
cigar, gff_fields['attrb']['Read'], ref_seq, vcf_pos,
var5p, var3p, parent_ini_pos, parent_end_pos, hairpin)
# Adding the variants to a dictionary and calculating all the fields of a vcf file format:
if len(key_var) > 0:
for s in range(len(key_var)):
key_dict = vcf_chrom + '-' + str(
key_pos[s]) + '-' + str(key_var[s])
raw_counts = gff_fields['attrb']['Expression']
raw_counts = [int(i) for i in raw_counts.split(',')]
nozero_counts = [
int(i > 0) for i in raw_counts
] # counts for every sample if expr != 0.
if gtf_name in mirna_dict: # Adding expression values to same mirnas
mirna_dict[gtf_name]['Z'] = [
sum(x) for x in zip(mirna_dict[gtf_name]['Z'],
raw_counts)
]
else:
mirna_dict[gtf_name] = {}
mirna_dict[gtf_name]["Z"] = raw_counts
if key_dict in all_dict:
if all_dict[key_dict]["Type"] in [
"A", "C", "T", "G"
]:
all_dict[key_dict]['X'] = [
sum(x) for x in zip(
all_dict[key_dict]['X'], nozero_counts)
]
all_dict[key_dict]['Y'] = [
sum(x) for x in zip(
all_dict[key_dict]['Y'], raw_counts)
]
else:
key_list.append(key_dict)
all_dict[key_dict] = {}
all_dict[key_dict]["Chrom"] = vcf_chrom
all_dict[key_dict]["Position"] = key_pos[s]
all_dict[key_dict]["mirna"] = gtf_name
all_dict[key_dict]["Type"] = key_var[s]
if key_var[s][0] in ["A", "C", "T", "G"]:
n_SNP += 1
all_dict[key_dict]["SNP"] = True
all_dict[key_dict]["ID"] = gff_fields['attrb'][
'Name'] + '-SNP' + str(n_SNP)
all_dict[key_dict]['X'] = nozero_counts
all_dict[key_dict]['Y'] = raw_counts
else:
n_noSNP += 1
all_dict[key_dict]["SNP"] = False
all_dict[key_dict]["ID"] = gff_fields['attrb'][
'Name'] + '-nonSNP' + str(n_noSNP)
all_dict[key_dict]["Ref"] = vcf_ref[s]
all_dict[key_dict]["Alt"] = vcf_alt[s]
all_dict[key_dict]["Qual"] = "."
all_dict[key_dict]["Filter"] = gff_fields['attrb'][
'Filter']
all_dict[key_dict]["Info"] = "NS=" + str(
len(sample_names))
else:
no_var += 1
# Writing the VCF file:
for s in key_list:
variant_line = (
"\n%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s" %
(all_dict[s]["Chrom"], all_dict[s]["Position"], all_dict[s]["ID"],
all_dict[s]["Ref"], all_dict[s]["Alt"], all_dict[s]["Qual"],
all_dict[s]["Filter"], all_dict[s]["Info"]))
if all_dict[s]["Type"] in ["A", "T", "C", "G"]:
format_col = "TRC:TSC:TMC:GT"
variant_line = variant_line + "\t" + format_col
samples = ""
for n in range(len(sample_names)):
X = all_dict[s]["X"][n]
Y = all_dict[s]["Y"][n]
Z = mirna_dict[all_dict[s]["mirna"]]["Z"][n]
# Calculating the genotype:
if Y == 0:
GT = "0|0"
elif Z == Y:
GT = "1|1"
else:
GT = "1|0"
samples = samples + "\t" + str(X) + ":" + str(Y) + ":" + str(
Z) + ":" + GT
variant_line = variant_line + samples
else:
format_col = ""
variant_line = variant_line + format_col
vcf_file.write(variant_line)
vcf_file.close()
def read_file(folder, args):
"""
Read sRNAbench file and convert to mirtop GFF format.
Args:
*fn(str)*: file name with sRNAbench output information.
*database(str)*: database name.
*args(namedtuple)*: arguments from command line.
See *mirtop.libs.parse.add_subparser_gff()*.
Returns:
*reads (nested dicts)*:gff_list has the format as
defined in *mirtop.gff.body.read()*.
"""
reads_anno = os.path.join(folder, "reads.annotation")
reads_iso = os.path.join(folder, "microRNAannotation.txt")
sep = " " if args.out_format == "gtf" else "="
sample = os.path.basename(folder)
database = args.database
precursors = args.precursors
matures = args.matures
n_out = 0
n_in = 0
n_ns = 0
n_notassign = 0
n_notindb = 0
reads = defaultdict(dict)
seen = set()
source_iso = _read_iso(reads_iso)
logger.info("Reads with isomiR information %s" % len(source_iso))
with open(reads_anno) as handle:
for sequence in handle:
cols = sequence.strip().split("\t")
query_name = cols[0]
query_sequence = cols[0]
if query_name not in reads and not query_sequence:
continue
if query_sequence and query_sequence.find("N") > -1:
n_ns += 1
continue
if cols[3].find("mature") == -1:
n_in += 1
continue
counts = int(cols[1])
hit = len(
set([mirna.split("#")[1] for mirna in cols[4].split("$")]))
for nhit in cols[4].split("$"):
logger.debug("SRNABENCH::line hit: %s" % nhit)
hit_info = nhit.split("#")
pos_info = hit_info[3].split(",")
start = int(pos_info[1]) - 1
end = start + len(query_sequence) # int(pos_info[2]) - 1
chrom = pos_info[0]
mirName = hit_info[1]
if chrom not in precursors or chrom not in matures:
n_notindb += 1
if mirName not in matures[chrom]:
n_notindb += 1
if (query_sequence, mirName) in seen:
continue
seen.add((query_sequence, mirName))
if (query_sequence, mirName) not in source_iso:
continue
isoformat = source_iso[(query_sequence, mirName)]
if isoformat == "mv":
n_notassign += 1
continue
source = "isomiR" if isoformat != "NA" else "ref_miRNA"
logger.debug("SRNABENCH::query: {query_sequence}\n"
" precursor {chrom}\n"
" name: {query_name}\n"
" start: {start}\n"
" external: {isoformat}\n"
" hit: {hit}".format(**locals()))
logger.debug("SRNABENCH:: start %s end %s" % (start, end))
if len(precursors[chrom]) < start + len(query_sequence):
n_out += 1
continue
Filter = "Pass"
cigar = make_cigar(query_sequence,
precursors[chrom][start:end])
preName = chrom
score = "."
strand = "+"
idu = make_id(query_sequence)
attrb = ("Read {query_sequence}; UID {idu}; Name {mirName};"
" Parent {preName}; Variant {isoformat};"
" Cigar {cigar}; Expression {counts};"
" Filter {Filter}; Hits {hit};").format(**locals())
line = ("{chrom}\t{database}\t{source}\t{start}\t{end}\t"
"{score}\t{strand}\t.\t{attrb}").format(**locals())
if args.add_extra:
extra = variant_with_nt(line, args.precursors,
args.matures)
line = "%s Changes %s;" % (line, extra)
line = paste_columns(read_gff_line(line), sep=sep)
if start not in reads[chrom]:
reads[chrom][start] = []
if Filter == "Pass":
n_in += 1
reads[chrom][start].append(
[idu, chrom, counts, sample, line])
logger.info("Loaded %s reads with %s hits" % (len(reads), n_in))
logger.info("Reads without precursor information: %s" % n_notindb)
logger.info("Reads with MV as variant definition,"
" not supported by GFF: %s" % n_notassign)
logger.info("Hit Filtered by having > 3 changes: %s" % n_out)
return reads
