def overlapping(self, release, chromosome, start, end, padding=0):
"""Return overlapping ``RegElement`` objects."""
return (super().all().filter(
release=release,
chromosome=chromosome,
bin__in=binning.overlapping_bins(start - 1, end),
start__lte=end + padding,
end__gte=max(0, start - padding),
))
def test_overlapping(intervals, interval):
start, stop = interval
# Intervals overlapping the query interval.
overlapping = set((x, y) for x, y in intervals
if x < stop and start < y)
# Pre-selection of intervals using binning.
binned = set((x, y) for x, y in intervals
if binning.assign_bin(x, y)
in binning.overlapping_bins(start, stop))
assert binned.issuperset(overlapping)
def chrom2c(self, variant, rt, gene=None):
"""
@arg variant: a variant description
@type variant: unicode
@arg rt: the return type
@type rt: unicode
@kwarg gene: Optional gene name. If given, return variant descriptions
on all transcripts for this gene.
@type gene: unicode
@return: HGVS_notatations ;
@rtype: dictionary or list
"""
if not self._parseInput(variant):
return None
acc = self.parseTree.LrgAcc or self.parseTree.RefSeqAcc
version = self.parseTree.Version
chromosome = Chromosome.query \
.filter_by(assembly=self.assembly,
accession='%s.%s' % (acc, version)).first()
if not chromosome:
self.__output.addMessage(
__file__, 4, "ENOTINDB",
"Accession number %s could not be found in our database or is "
"not suitable for the requested conversion." % acc)
return None
#if
if self.parseTree.SingleAlleleVarSet:
variants = [v.RawVar for v in self.parseTree.SingleAlleleVarSet]
else:
variants = [self.parseTree.RawVar]
min_loc = 9000000000
max_loc = 0
for variant in variants:
#FIXME This should be a proper conversion.
loc = int(variant.StartLoc.PtLoc.Main)
if variant.EndLoc:
loc2 = int(variant.EndLoc.PtLoc.Main)
else:
loc2 = loc
if loc2 < loc:
self.__output.addMessage(
__file__, 3, 'ERANGE', 'End position is '
'smaller than the begin position.')
return None
min_loc = min(min_loc, loc)
max_loc = max(max_loc, loc2)
if gene:
mappings = chromosome.transcript_mappings.filter_by(gene=gene)
else:
start = max(min_loc - 5000, 1)
stop = min(max_loc + 5000, binning.MAX_POSITION + 1)
bins = binning.overlapping_bins(start - 1, stop)
mappings = chromosome.transcript_mappings.filter(
TranscriptMapping.bin.in_(bins),
TranscriptMapping.start <= stop,
TranscriptMapping.stop >= start).order_by(
TranscriptMapping.start, TranscriptMapping.stop,
TranscriptMapping.gene, TranscriptMapping.accession,
TranscriptMapping.version, TranscriptMapping.transcript)
HGVS_notatations = defaultdict(list)
NM_list = []
for mapping in mappings:
self._reset()
self.mapping = mapping
core_mapping = self._coreMapping()
if not core_mapping:
#balen
continue
reference = self.mapping.reference
geneName = self.mapping.gene
strand = self.mapping.orientation == 'forward'
# Check if n or c type
# Note: Originally, the below check using crossmap.info() was
# used (commented out now), but I do not understand this
# logic. Also, it breaks n. notation on non-coding mtDNA
# transcripts, so I replaced it with a simple .CDS check.
#info = self.crossmap.info()
#if info[0] == 1 and info[1] == info[2] :
# mtype = 'n'
#else :
# mtype = 'c'
if self.crossmap.CDS:
mtype = 'c'
else:
mtype = 'n'
mutations = []
for variant, cmap in zip(variants, core_mapping):
try:
f_change = _construct_change(variant)
r_change = _construct_change(variant, reverse=True)
except NotImplementedError as e:
self.__output.addMessage(__file__,
4, "ENOTIMPLEMENTEDERROR",
unicode(e))
return None
startp = self.crossmap.tuple2string(
(cmap.startmain, cmap.startoffset))
endp = self.crossmap.tuple2string(
(cmap.endmain, cmap.endoffset))
if strand:
change = f_change
else:
change = r_change
startp, endp = endp, startp
if cmap.start_g != cmap.end_g:
loca = "%s_%s" % (startp, endp)
else:
loca = "%s" % startp
mutations.append('%s%s' % (loca, change))
if len(mutations) == 1:
mutation = mutations[0]
else:
mutation = '[' + ';'.join(mutations) + ']'
description = "%s:%c.%s" % (reference, mtype, mutation)
HGVS_notatations[geneName].append(description)
NM_list.append(description)
#for
if rt == "list":
return NM_list
return HGVS_notatations
def chrom2c(self, variant, rt, gene=None):
"""
@arg variant: a variant description
@type variant: unicode
@arg rt: the return type
@type rt: unicode
@kwarg gene: Optional gene name. If given, return variant descriptions
on all transcripts for this gene.
@type gene: unicode
@return: HGVS_notatations ;
@rtype: dictionary or list
"""
if not self._parseInput(variant) :
return None
acc = self.parseTree.LrgAcc or self.parseTree.RefSeqAcc
version = self.parseTree.Version
chromosome = Chromosome.query \
.filter_by(assembly=self.assembly,
accession='%s.%s' % (acc, version)).first()
if not chromosome :
self.__output.addMessage(__file__, 4, "ENOTINDB",
"Accession number %s could not be found in our database or is "
"not suitable for the requested conversion." %
acc)
return None
#if
if self.parseTree.SingleAlleleVarSet:
variants = [v.RawVar for v in self.parseTree.SingleAlleleVarSet]
else:
variants = [self.parseTree.RawVar]
min_loc = 9000000000
max_loc = 0
for variant in variants:
#FIXME This should be a proper conversion.
loc = int(variant.StartLoc.PtLoc.Main)
if variant.EndLoc :
loc2 = int(variant.EndLoc.PtLoc.Main)
else :
loc2 = loc
if loc2 < loc:
self.__output.addMessage(__file__, 3, 'ERANGE', 'End position is '
'smaller than the begin position.')
return None
min_loc = min(min_loc, loc)
max_loc = max(max_loc, loc2)
if gene:
mappings = chromosome.transcript_mappings.filter_by(gene=gene)
else:
start = max(min_loc - 5000, 1)
stop = min(max_loc + 5000, binning.MAX_POSITION + 1)
bins = binning.overlapping_bins(start - 1, stop)
mappings = chromosome.transcript_mappings.filter(
TranscriptMapping.bin.in_(bins),
TranscriptMapping.start <= stop,
TranscriptMapping.stop >= start
).order_by(
TranscriptMapping.start,
TranscriptMapping.stop,
TranscriptMapping.gene,
TranscriptMapping.accession,
TranscriptMapping.version,
TranscriptMapping.transcript)
HGVS_notatations = defaultdict(list)
NM_list = []
for mapping in mappings:
self._reset()
self.mapping = mapping
core_mapping = self._coreMapping()
if not core_mapping:
#balen
continue
reference = self.mapping.reference
geneName = self.mapping.gene
strand = self.mapping.orientation == 'forward'
# Check if n or c type
# Note: Originally, the below check using crossmap.info() was
# used (commented out now), but I do not understand this
# logic. Also, it breaks n. notation on non-coding mtDNA
# transcripts, so I replaced it with a simple .CDS check.
#info = self.crossmap.info()
#if info[0] == 1 and info[1] == info[2] :
# mtype = 'n'
#else :
# mtype = 'c'
if self.crossmap.CDS:
mtype = 'c'
else:
mtype = 'n'
mutations = []
for variant, cmap in zip(variants, core_mapping):
try:
f_change = _construct_change(variant)
r_change = _construct_change(variant, reverse=True)
except NotImplementedError as e:
self.__output.addMessage(__file__, 4,
"ENOTIMPLEMENTEDERROR", unicode(e))
return None
startp = self.crossmap.tuple2string((cmap.startmain, cmap.startoffset))
endp = self.crossmap.tuple2string((cmap.endmain, cmap.endoffset))
if strand :
change = f_change
else :
change = r_change
startp, endp = endp, startp
if cmap.start_g != cmap.end_g :
loca = "%s_%s" % (startp, endp)
else :
loca = "%s" % startp
mutations.append('%s%s' % (loca, change))
if len(mutations) == 1:
mutation = mutations[0]
else:
mutation = '[' + ';'.join(mutations) + ']'
description = "%s:%c.%s" % (reference, mtype, mutation)
HGVS_notatations[geneName].append(description)
NM_list.append(description)
#for
if rt == "list" :
return NM_list
return HGVS_notatations
def test_overlapping_bins(start, stop, expected):
assert binning.overlapping_bins(start, stop) == expected
