class TestGffWriter:
RECORD1 = Gff3Record("chr1",
10,
11,
"insertion",
attributes=[("cat", "1"), ("dog", "2")])
RECORD2 = Gff3Record("chr1",
200,
201,
"substitution",
attributes=[("mouse", "1"), ("moose", "2")])
def setup_method(self):
self.outfile = StringIO()
self.gffWriter = GffWriter(self.outfile)
def test_writeHeader(self):
self.gffWriter.writeHeader("##foo bar")
assert "##gff-version 3\n##foo bar\n" == self.outfile.getvalue()
def test_writeRecord(self):
self.gffWriter.writeRecord(self.RECORD1)
self.gffWriter.writeRecord(self.RECORD2)
expected = (
"##gff-version 3\n" +
"chr1\t.\tinsertion\t10\t11\t.\t.\t.\tcat=1;dog=2\n" +
"chr1\t.\tsubstitution\t200\t201\t.\t.\t.\tmouse=1;moose=2\n")
assert expected == self.outfile.getvalue()
def setup(self):
self.outfile = StringIO()
self.record1 = Gff3Record("chr1", 10, 11, "insertion",
attributes=[("cat", "1"), ("dog", "2")])
self.record2 = Gff3Record("chr1", 200, 201, "substitution",
attributes=[("mouse", "1"), ("moose", "2")])
self.gffWriter = GffWriter(self.outfile)
def run(argv):
nuc = []
basemods = []
while len(nuc) < 5000:
x = random.random()
if x > 0.99 and len(nuc) < 4880:
n = len(nuc) + 3
basemods.append(
Gff3Record("genome", n, n, "modified_base", 100, "+", ".",
"kinModCall", [("coverage", "100"),
("IPDRatio", "4.0"),
("identificationQv", "50")]))
nuc.extend(["G", "T", "A", "C"])
elif x < 0.01 and len(nuc) < 4880:
n = len(nuc) + 3
basemods.append(
Gff3Record("genome", n, n, "modified_base", 100, "+", ".",
"kinModCall", [("coverage", "100"),
("IPDRatio", "4.0"),
("identificationQv", "50")]))
nuc.extend(["C", "T", "A", "G"])
else:
b = "ACGT"[random.randint(0, 3)]
if ((b == "C" and nuc[-3:] == ["G", "T", "A"])
or (b == "G" and nuc[-3:] == ["C", "T", "A"])):
nuc.append("T")
else:
nuc.append(b)
seq = "".join(nuc)
with FastaWriter("genome.fasta") as fa_out:
fa_out.writeRecord("genome", "".join(nuc))
with GffWriter("basemods.gff") as gff_out:
for rec in basemods:
ctx_start = max(0, rec.start - 21)
ctx_end = min(len(seq), rec.end + 20)
context = seq[ctx_start:ctx_end]
rec.attributes["context"] = context
gff_out.writeRecord(rec)
return 0
class TestGff3Record:
RECORD = Gff3Record("chr1",
10,
11,
"insertion",
attributes=[("cat", "1"), ("dog", "2")])
def test_str(self):
assert "chr1\t.\tinsertion\t10\t11\t.\t.\t.\tcat=1;dog=2" == str(
self.RECORD)
def test_modification(self):
record = self.RECORD.copy()
record.dog = 3
record.cat = 4
record.mouse = 5
record.start = 100
record.end = 110
assert "chr1\t.\tinsertion\t100\t110\t.\t.\t.\tcat=4;dog=3;mouse=5" == str(
record)
def test_fromString(self):
newRecord = Gff3Record.fromString(str(self.RECORD))
assert str(self.RECORD) == str(newRecord)
def test_get(self):
"""
Verify field access behavior
"""
record = self.RECORD
record.dog = 3
record.cat = 4
record.mouse = 5
record.start = 100
record.end = 110
assert 3 == record.dog
assert 100 == record.start
with pytest.raises(AttributeError):
record.god
assert 3 == record.get("dog")
assert record.get("god") is None
assert 100 == record.get("start", 100)
def toGffRecord(var):
varType = var.variantType
gffType = varType.lower()
gffStart = (var.refStart + 1) if (var.refSeq != "") else var.refStart
gffEnd = var.refEnd if (var.refSeq != "") else var.refStart
gffFreq = gffVariantFrequency(var)
record = Gff3Record(reference.idToFullName(var.refId), gffStart, gffEnd, gffType)
record.reference = var.refSeq or "."
record.variantSeq = gffVariantSeq(var)
if gffFreq:
record.frequency = gffFreq
record.coverage = var.coverage
record.confidence = var.confidence
if var.annotations:
for (k, v) in var.annotations:
record.put(k, v)
return record
def makeM5CgffRecord(self, siteObs):
start = siteObs['tpl'] + 1
end = siteObs['tpl'] + 1
attributes = [('coverage', siteObs['coverage']),
('IPDRatio', siteObs['ipdRatio'])]
recordType = 'CG'
refName = siteObs['refId']
score = "%.3f" % siteObs['Ca5C']
strand = '+' if siteObs['strand'] == 0 else '-'
return Gff3Record(refName, start, end,
type=recordType,
score=score,
strand=strand,
source='kinModCall',attributes=attributes)
def makeGffRecord(self, siteObs):
"""
Convert the internal site observation object into a GFF entry
"""
# Some useful attributes about the observation
# - cognate base
# - context snippet
# - ipd ratio
# - coverage
snippet = self.snippetFunc(siteObs['tpl'], siteObs['strand'])
attributes = [('coverage', siteObs['coverage']),
('context', snippet),
('IPDRatio', siteObs['ipdRatio'])]
# Base of detected mod -- single position, closed,open
# interval.
# Note -- internally the tool uses 0-based reference
# coordinates, however in gff the template indices are
# 1-based. Make that adjustment here.
# On start vs. end: My reading of the gff spec
# (http://www.sequenceontology.org/resources/gff3.html) says
# to me that 1-base long feature (e.g. a modified base) should
# have start + 1 == end, and 0-base long features
# (e.g. insertions) should have start == end. This is not the
# convention that Marco has apdopted in SMRTView, or the
# convention that EviCons originally used. We will adopt
# their convention here, for now.
start = siteObs['tpl'] + 1
end = siteObs['tpl'] + 1
if siteObs.has_key('motif'):
attributes.append(('motif', "%s" % siteObs['motif']))
if siteObs.has_key('id'):
attributes.append(('id', "%s" % siteObs['id']))
if self.options.methylFraction and siteObs.has_key(FRAC):
attributes.append(('frac', "%.3f" % siteObs[FRAC]))
attributes.append(('fracLow', "%.3f" % siteObs[FRAClow]))
attributes.append(('fracUp', "%.3f" % siteObs[FRACup]))
if siteObs.has_key('modificationScore'):
# Report the QV from the modification identification module as a special tag
attributes.append(('identificationQv', "%d" % int(round(siteObs['modificationScore']))))
if siteObs.has_key('modification'):
if siteObs['modification'] == '.':
recordType = 'modified_base'
elif siteObs['modification'] == 'nMd':
recordType = '.'
else:
# if we have an identified mod, use it; otherwise use the old generic term
recordType = siteObs['modification']
else:
recordType = 'modified_base'
refName = siteObs['refName']
score = int(round(siteObs['score']))
strand = '+' if siteObs['strand'] == 0 else '-'
return Gff3Record(refName, start, end,
type=recordType,
score=score,
strand=strand,
source='kinModCall',
attributes=attributes)
return rec
def setup(self):
self.record = Gff3Record("chr1",
10,
11,
"insertion",
attributes=[("cat", "1"), ("dog", "2")])
