def test_updateFeatureGff(self):
iesgff = SharedFunctions.Gff()
iesgff.list2gff(TestInsert.gfflist)
ins = Insert.Insert(TestInsert.ref, iesgff, TestInsert.ies)
ins._filterInserts()
ins._updatePositionsInserts()
feats = [i.split("\t") for i in TestInsert.oldfeatures]
newgff = ins.updateFeatureGff(feats, addsuffix=True)
self.assertEqual(
[[str(elem) for elem in i] for i in newgff if re.match('ID=gene1.', i[8])],
[['ctg1','.','gene','3','5','.','.','.','ID=gene1.seg_0;key1=attr1;key2=attr2'],
['ctg1','.','gene','10','11','.','.','.','ID=gene1.seg_1;key1=attr1;key2=attr2']])
self.assertEqual(
[[str(elem) for elem in i] for i in newgff if re.match('ID=gene2', i[8])],
[['ctg1','.','gene','21','24','.','.','.','ID=gene2']])
# Test multi-segment feature spanning several lines
self.assertEqual(
[[str(elem) for elem in i] for i in newgff if re.match('ID=cds1', i[8])],
[['ctg1','.','CDS','3','5','.','+','0','ID=cds1.seg_0;key1=attr1;key2=attr2'],
['ctg1','.','CDS','10','11','.','+','0','ID=cds1.seg_1;key1=attr1;key2=attr2'],
['ctg1','.','CDS','21','24','.','+','0','ID=cds1']])
# Test feature without added suffix
newgff_nosuffix = ins.updateFeatureGff(feats, addsuffix=False)
self.assertEqual(
[[str(elem) for elem in i] for i in newgff_nosuffix if re.match('ID=gene1;', i[8])],
[['ctg1','.','gene','3','5','.','.','.','ID=gene1;key1=attr1;key2=attr2'],
['ctg1','.','gene','10','11','.','.','.','ID=gene1;key1=attr1;key2=attr2']])
def test_reportInsertedReference(self):
gff = SharedFunctions.Gff()
gff.list2gff(TestInsert.gfflist)
ins = Insert.Insert(TestInsert.ref, gff, TestInsert.ies)
newfasta, newgff = ins.reportInsertedReference()
self.assertEqual(
str(newfasta['ctg1'].seq),
'AAAAATTTTAAAAGGGGGAAAAAAAAAAA')
self.assertEqual(
str(newfasta['ctg2'].seq),
'GGGGGGGGGCCCCCGGGGGGGGGGG')
def insert(args):
logger = logging.getLogger("main.insert")
run_boilerplate(logger)
logger.info("Started BleTIES Insert")
logger.info("Reading input files")
refgenome = SeqIO.to_dict(SeqIO.parse(args.ref, "fasta"))
gff = SharedFunctions.Gff()
gff.file2gff(args.ies)
if re.match(r"ins", args.mode):
ies = SeqIO.to_dict(SeqIO.parse(args.iesfasta, "fasta"))
logger.info(
"Inserting IESs to MAC reference to make MAC+IES hybrid reference")
ins = Insert.Insert(refgenome, gff, ies)
newrefgenome, newgff = ins.reportInsertedReference()
if args.featuregff:
logger.info(f"Updating coords in feature table {args.featuregff}")
# Slurp file into memory as list of lists
with open(args.featuregff, 'r') as fh:
feats = [
line.rstrip().split("\t") for line in fh
if line[0] != '#' and len(line.rstrip().split("\t")) == 9
]
# Update coordinates of feature table to account for inserted IESs
newfeaturegff = ins.updateFeatureGff(feats, args.addsuffix)
outfeaturegff = f"{args.out}.iesplus.feature_table.gff"
logger.info(f"Writing new feature table to {outfeaturegff}")
with open(outfeaturegff, 'w') as fh:
fh.write("##gff-version 3\n")
for line in newfeaturegff:
fh.write("\t".join([str(i) for i in line]))
fh.write("\n")
outfasta = f"{args.out}.iesplus.fasta"
outgff = f"{args.out}.iesplus.gff"
elif re.match(r"del", args.mode):
logger.info(
f"Removing IESs from MAC+IES reference to make MAC-IES reference")
dels = Insert.Insert(refgenome, gff, None)
newrefgenome, newgff = dels.reportDeletedReference()
outfasta = f"{args.out}.iesminus.fasta"
outgff = f"{args.out}.iesminus.gff"
logger.info(f"Writing output files to {outfasta}, {outgff}")
with open(outfasta, "w") as fh:
SeqIO.write(list(newrefgenome.values()), fh, "fasta")
newgff.gff2file(outgff, header=True)
logger.info("Finished Insert")
def test_updateFeatureGff_tapointer(self):
iesgff = SharedFunctions.Gff()
iesgff.list2gff(TestInsert.gfflist)
ins = Insert.Insert(TestInsert.ref, iesgff, TestInsert.ies)
ins._filterInserts()
ins._updatePositionsInserts()
ins._updatePointerPositionsInserts()
ins._addSequences()
# print("\n".join(ins._newgff.gff2list())) # testing
self.assertEqual(str(ins._newgff.getValue('ies6', 'start')), '15')
self.assertEqual(str(ins._newgff.getValue('ies6', 'end')), '19')
self.assertEqual(str(ins._newgff.getAttr('ies6', 'ta_pointer_start')), '16')
self.assertEqual(str(ins._newgff.getAttr('ies6', 'ta_pointer_end')), '20')
self.assertEqual(str(ins._newgff.getValue('ies7','start')), '11')
self.assertEqual(str(ins._newgff.getValue('ies7','end')), '15')
self.assertEqual(str(ins._newgff.getAttr('ies7','ta_pointer_start')), '11')
self.assertEqual(str(ins._newgff.getAttr('ies7','ta_pointer_end')), '15')
def test_reportDeletedReference(self):
gff = SharedFunctions.Gff()
gff.list2gff(TestInsert.gfflist)
# Insert sequences into reference
ins = Insert.Insert(TestInsert.ref, gff, TestInsert.ies)
newfasta, newgff = ins.reportInsertedReference()
# print("\n".join(newgff.gff2list())) # testing
# Take them out again
dels = Insert.Insert(newfasta, newgff, None)
delfasta, delgff = dels.reportDeletedReference()
# print("\n".join(delgff.gff2list())) # testing
# Check that they are the same sequence
self.assertEqual(
str(delfasta['ctg1'].seq),
str(TestInsert.ref['ctg1'].seq))
self.assertEqual(
str(delfasta['ctg2'].seq),
str(TestInsert.ref['ctg2'].seq))
self.assertEqual(
str(delfasta['ctg3'].seq),
str(TestInsert.ref['ctg3'].seq))
self.assertEqual(str(delgff.getAttr('ies6', 'ta_pointer_start')), '10')
self.assertEqual(str(delgff.getAttr('ies6', 'ta_pointer_end')), '10')
type=int,
help="Minimum length to show in histogram of IES lengths (detail)")
parser.add_argument(
"--hist_len_max",
default=400,
type=int,
help="Maximum length to show in histogram of IES lengths (detail)")
parser.add_argument(
"--hist_style",
default="facet",
help="Style for histograms of IES lengths: 'facet', 'bar', or 'barstacked'"
)
args = parser.parse_args()
# Import GFF records
gff_obj = SharedFunctions.Gff()
gff_obj.file2gff(args.gff)
# Get relevant fields
out = []
for gff_id in gff_obj:
row = {}
row['id'] = gff_id
if int(gff_obj.getValue(gff_id,
"start")) == int(gff_obj.getValue(gff_id, "end")):
row['type'] = "ins"
elif int(gff_obj.getValue(gff_id, "start")) < int(
gff_obj.getValue(gff_id, "end")):
row['type'] = "del"
else:
raise Exception(f"Invalid GFF3, start > end: check {row['id']}")
def miser(args):
logger = logging.getLogger("main.miser")
run_boilerplate(logger)
logger.info("Started BleTIES MISER")
# Read input files
alnfile, refgenome = read_bam_ref(args.bam, args.ref)
# Initialize new IesRecords object to store putative IESs
iesrecords = Milraa.IesRecords(alnfile, "bam", refgenome)
# Read in IES GFF file produced by Milraa
logger.info("Reading GFF file containing putative IESs")
iesgff = SharedFunctions.Gff()
iesgff.file2gff(args.gff)
# Compare mean mismatch percentage of reads with and without putative IES
# for each putative IES
logger.info("""
Reporting possibly spurious IESs due to misassembly or mapped paralogs
""")
out_gff_split = defaultdict(
list) # dict to hold split GFF file keyed by diagnosis
args.out.write("\t".join([
'ID', 'mean_mismatch_pc_with_indel', 'mean_mismatch_pc_no_indel',
'stdev_with_indel', 'stdev_no_indel', 'statistic', 'p-value',
'num_reads_with_indel', 'num_reads_no_indel', 'diagnosis'
]))
args.out.write("\n")
for bpid in iesgff:
ins_mm, non_mm = iesrecords.reportIndelReadMismatchPc(
iesgff.getValue(bpid, 'seqid'), int(iesgff.getValue(bpid,
'start')),
int(iesgff.getValue(bpid, 'end')),
int(iesgff.getAttr(bpid, 'IES_length')), int(args.min_ies_length))
# Perform test of mismatch % if more than 2 reads with inserts
# (otherwise stdev meaningless)
if len(ins_mm) > 2 and len(non_mm) > 2:
if args.spurious_ies_test == 'mann-whitney':
# Mann-Whitney U test for whether mismatch % with indel of interest
# is greater than without
mwstat, mwpval = mannwhitneyu(ins_mm,
non_mm,
alternative='greater')
else:
# Ward's t-test (non-equal population variances)
mwstat, mwpval = ttest_ind(ins_mm, non_mm, equal_var=False)
# Report
outarr = [
bpid,
round(stats.mean(ins_mm), 2),
round(stats.mean(non_mm), 2),
round(stats.stdev(ins_mm), 2),
round(stats.stdev(non_mm), 2),
round(mwstat, 2),
'%.2E' % mwpval, # scientific notation
len(ins_mm),
len(non_mm)
]
# Diagnosis
diagnosis = "ok"
if stats.mean(ins_mm) > 5 or stats.mean(non_mm) > 5:
diagnosis = "high_error"
if len(ins_mm) > len(non_mm):
diagnosis = 'misassembly'
# PVAL_UNCORR = 0.05 # TODO magic number
pval_corr = args.spurious_ies_pvalue / \
len(iesgff) # Bonferroni correction
if mwpval < pval_corr and stats.mean(ins_mm) > stats.mean(non_mm):
diagnosis = "paralog"
elif len(non_mm) < 1:
outarr = [
bpid,
round(stats.mean(ins_mm), 2), "NA", "NA", "NA", "NA", "NA",
len(ins_mm),
len(non_mm)
]
diagnosis = "misassembly" # or scrambling
else:
outarr = [
bpid,
round(stats.mean(ins_mm), 2),
round(stats.mean(non_mm), 2), "NA", "NA", "NA", "NA",
len(ins_mm),
len(non_mm)
]
diagnosis = "low_coverage"
outarr.append(diagnosis)
# Split the input GFF entries into each diagnosis group
out_gff_split[diagnosis].append(iesgff.getEntry(bpid))
# Write output
args.out.write("\t".join([str(i) for i in outarr]))
args.out.write("\n")
# args.out.write(" ".join([str(i) for i in ins_mm]) + "\n")
# args.out.write(" ".join([str(i) for i in non_mm]) + "\n")
# Output split GFF files
if args.split_gff:
logger.info("Splitting input GFF entries into inferred categories")
for diag in out_gff_split:
outfile = f"{args.gff}.{diag}.gff3"
with open(outfile, "w") as fh_spl:
# Write gff version header and comment some info on this file
fh_spl.write("##gff-version 3\n")
fh_spl.write("# " + " ".join(sys.argv) + "\n")
fh_spl.write(
f"# BleTIES MISER putative IESs classified as {diag}\n")
for line in out_gff_split[diag]:
fh_spl.write("\t".join([str(i) for i in line]))
fh_spl.write("\n")
# Close alignment filehandle
alnfile.close()
logger.info("Finished MISER")