def products(self):
if self._products:
return self._products
self._products = []
for fp in self.fwd_primers:
for rp in self.rev_primers:
if self.template.circular and fp.position>rp.position:
tmpl = self.template.shifted(fp.position-len(fp.footprint))
tmpl = tmpl._multiply_circular(2)
tmpl = tmpl[:len(self.template) - (fp.position - rp.position) + len(rp.footprint) + len(fp.footprint)]
#print len(self.template) - (fp.position - rp.position) + len(rp.footprint) + len(fp.footprint)
#print len(tmpl)
elif self.template.circular:
tmpl = self.template._multiply_circular(3)
tmpl = tmpl[fp.position-len(fp.footprint)+len(self.template):rp.position+len(rp.footprint)+len(self.template)]
else:
tmpl = self.template[fp.position-len(fp.footprint):rp.position+len(rp.footprint)]
prd = ( Dseqrecord(fp.tail) + tmpl + Dseqrecord(rp.tail).reverse_complement())
prd.add_feature( 0, len(fp), label=fp.id)
prd.add_feature( len(prd)-len(rp),len(prd),label=rp.id, strand=-1)
#prd.seq = fp.seq+tmpl.seq[len(fp.footprint):len(tmpl)-len(rp.footprint)]+rp.seq.reverse_complement()
#features = tmpl[fp.position-len(fp.footprint):rp.position+len(rp.footprint)].features
#print fp.position-len(fp.footprint), rp.position+len(rp.footprint), features
#print "<<<<<<<<<<<<<<",features
#prd.features = [f._shift(len(fp.tail)) for f in features]
# description = Genbank LOCUS max 16 chars
prd.name = "{0}bp_PCR_prod".format(len(prd))[:16]
prd.id = "{0}bp {1}".format( str(len(prd))[:14], prd.seguid() )
prd.description="Product_{0}_{1}".format( fp.description,
rp.description)
self._products.append( Amplicon(prd,
template=tmpl,
forward_primer=fp,
reverse_primer=rp,
saltc=50,
forward_primer_concentration=1000,
reverse_primer_concentration=1000))
assert " " not in str(prd.seq.watson)
assert " " not in str(prd.seq.crick)
return self._products
def _assemble(self):
for dr in self.dsrecs:
if dr.name in ("", ".", "<unknown name>", None):
dr.name = "frag{}".format(len(dr))
if self.only_terminal_overlaps:
algorithm = terminal_overlap
else:
algorithm = common_sub_strings
# analyze_overlaps
cols = {}
for dsrec in self.dsrecs:
dsrec.features = [f for f in dsrec.features if f.type != "overlap"]
dsrec.seq = Dseq(dsrec.seq.todata)
rcs = {dsrec: dsrec.rc() for dsrec in self.dsrecs}
matches = []
dsset = OrderedSet()
for a, b in itertools.combinations(self.dsrecs, 2):
match = algorithm(
str(a.seq).upper(),
str(b.seq).upper(), self.limit)
if match:
matches.append((a, b, match))
dsset.add(a)
dsset.add(b)
match = algorithm(
str(a.seq).upper(),
str(rcs[b].seq).upper(), self.limit)
if match:
matches.append((a, rcs[b], match))
dsset.add(a)
dsset.add(rcs[b])
matches.append(
(rcs[a], b, [(len(a) - sa - le, len(b) - sb - le, le)
for sa, sb, le in match]))
dsset.add(b)
dsset.add(rcs[a])
self.no_of_olaps = 0
for a, b, match in matches:
for start_in_a, start_in_b, length in match:
self.no_of_olaps += 1
chksum = a[start_in_a:start_in_a + length].seguid()
#assert chksum == b[start_in_b:start_in_b+length].seguid()
try:
fcol, revcol = cols[chksum]
except KeyError:
fcol = '#%02X%02X%02X' % (random.randint(
175, 255), random.randint(
175, 255), random.randint(175, 255))
rcol = '#%02X%02X%02X' % (random.randint(
175, 255), random.randint(
175, 255), random.randint(175, 255))
cols[chksum] = fcol, rcol
qual = {
"note": ["olp_{}".format(chksum)],
"chksum": [chksum],
"ApEinfo_fwdcolor": [fcol],
"ApEinfo_revcolor": [rcol]
}
if not chksum in [
f.qualifiers["chksum"][0]
for f in a.features if f.type == "overlap"
]:
a.features.append(
SeqFeature(FeatureLocation(start_in_a,
start_in_a + length),
type="overlap",
qualifiers=qual))
if not chksum in [
f.qualifiers["chksum"][0]
for f in b.features if f.type == "overlap"
]:
b.features.append(
SeqFeature(FeatureLocation(start_in_b,
start_in_b + length),
type="overlap",
qualifiers=qual))
for ds in dsset:
ds.features = sorted([f for f in ds.features],
key=operator.attrgetter("location.start"))
self.analyzed_dsrecs = list(dsset)
# Create graph
self.G = nx.MultiDiGraph(multiedges=True,
name="original graph",
selfloops=False)
self.G.add_node('5')
self.G.add_node('3')
for i, dsrec in enumerate(self.analyzed_dsrecs):
overlaps = sorted({
f.qualifiers['chksum'][0]: f
for f in dsrec.features if f.type == 'overlap'
}.values(),
key=operator.attrgetter('location.start'))
if overlaps:
overlaps = ([
SeqFeature(FeatureLocation(0, 0),
type='overlap',
qualifiers={'chksum': ['5']})
] + overlaps + [
SeqFeature(FeatureLocation(len(dsrec), len(dsrec)),
type='overlap',
qualifiers={'chksum': ['3']})
])
for olp1, olp2 in itertools.combinations(overlaps, 2):
n1 = olp1.qualifiers['chksum'][0]
n2 = olp2.qualifiers['chksum'][0]
if n1 == '5' and n2 == '3':
continue
s1, e1, s2, e2 = (
olp1.location.start.position,
olp1.location.end.position,
olp2.location.start.position,
olp2.location.end.position,
)
source_fragment = Fragment(dsrec, s1, e1, s2, e2, i)
self.G.add_edge(n1,
n2,
frag=source_fragment,
weight=s1 - e1,
i=i)
#linear assembly
linear_products = defaultdict(list)
for path in all_simple_paths_edges(self.G,
'5',
'3',
data=True,
cutoff=self.max_nodes):
pred_frag = copy(path[0][2].values().pop()['frag'])
source_fragments = [
pred_frag,
]
if pred_frag.start2 < pred_frag.end1:
result = pred_frag[pred_frag.start2 +
(pred_frag.end1 -
pred_frag.start2):pred_frag.end2]
else:
result = pred_frag[pred_frag.end1:pred_frag.end2]
for first_node, second_node, edgedict in path[1:]:
edgedict = edgedict.values().pop()
f = copy(edgedict['frag'])
f.alignment = pred_frag.alignment + pred_frag.start2 - f.start1
source_fragments.append(f)
if f.start2 > f.end1:
result += f[f.end1:f.end2]
else:
result += f[f.start2 + (f.end1 - f.start2):f.end2]
pred_frag = f
add = True
for lp in linear_products[len(result)]:
if (str(result.seq).lower() == str(lp.seq).lower()
or str(result.seq).lower() == str(
lp.seq.reverse_complement()).lower()):
add = False
for dsrec in self.dsrecs:
if (str(result.seq).lower() == str(dsrec.seq).lower()
or str(result.seq).lower() == str(
dsrec.seq.reverse_complement()).lower()):
add = False
if add:
linear_products[len(result)].append(
Contig(result, source_fragments))
self.linear_products = list(
itertools.chain.from_iterable(
linear_products[size]
for size in sorted(linear_products, reverse=True)))
# circular assembly
self.cG = self.G.copy()
self.cG.remove_nodes_from(('5', '3'))
#circular_products=defaultdict(list)
circular_products = {}
for pth in all_circular_paths_edges(self.cG):
ns = min(enumerate(pth), key=lambda x: x[1][2]['i'])[0]
path = pth[ns:] + pth[:ns]
pred_frag = copy(path[0][2]['frag'])
source_fragments = [
pred_frag,
]
if pred_frag.start2 < pred_frag.end1:
result = pred_frag[pred_frag.start2 +
(pred_frag.end1 -
pred_frag.start2):pred_frag.end2]
else:
result = pred_frag[pred_frag.end1:pred_frag.end2]
result.seq = Dseq(str(result.seq))
for first_node, second_node, edgedict in path[1:]:
f = copy(edgedict['frag'])
f.alignment = pred_frag.alignment + pred_frag.start2 - f.start1
source_fragments.append(f)
if f.start2 > f.end1:
nxt = f[f.end1:f.end2]
else:
nxt = f[f.start2 + (f.end1 - f.start2):f.end2]
nxt.seq = Dseq(str(nxt.seq))
result += nxt
pred_frag = f
#add=True
#for cp in circular_products[len(result)]:
# if (str(result.seq).lower() in str(cp.seq).lower()*2
# or
# str(result.seq).lower() == str(cp.seq.reverse_complement()).lower()*2):
# pass
# add=False
# print "##--"
#if add:
# circular_products[len(result)].append( Contig( Dseqrecord(result, circular=True), source_fragments))
r = Dseqrecord(result, circular=True)
circular_products[r.cseguid()] = Contig(r, source_fragments)
#self.circular_products = list(itertools.chain.from_iterable(circular_products[size] for size in sorted(circular_products, reverse=True)))
self.circular_products = sorted(circular_products.values(),
key=len,
reverse=True)
def _assemble(self):
for dr in self.dsrecs:
if dr.name in ("",".", "<unknown name>", None):
dr.name = "frag{}".format(len(dr))
if self.only_terminal_overlaps:
algorithm = terminal_overlap
else:
algorithm = common_sub_strings
# analyze_overlaps
cols = {}
for dsrec in self.dsrecs:
dsrec.features = [f for f in dsrec.features if f.type!="overlap"]
dsrec.seq = Dseq(dsrec.seq.todata)
rcs = {dsrec:dsrec.rc() for dsrec in self.dsrecs}
matches=[]
dsset=OrderedSet()
for a, b in itertools.combinations(self.dsrecs, 2):
match = algorithm( str(a.seq).upper(),
str(b.seq).upper(),
self.limit)
if match:
matches.append((a, b, match))
dsset.add(a)
dsset.add(b)
match = algorithm( str(a.seq).upper(),
str(rcs[b].seq).upper(),
self.limit)
if match:
matches.append((a, rcs[b], match))
dsset.add(a)
dsset.add(rcs[b])
matches.append((rcs[a], b, [(len(a)-sa-le,len(b)-sb-le,le) for sa,sb,le in match]))
dsset.add(b)
dsset.add(rcs[a])
self.no_of_olaps=0
for a, b, match in matches:
for start_in_a, start_in_b, length in match:
self.no_of_olaps+=1
chksum = a[start_in_a:start_in_a+length].seguid()
#assert chksum == b[start_in_b:start_in_b+length].seguid()
try:
fcol, revcol = cols[chksum]
except KeyError:
fcol = '#%02X%02X%02X' % (random.randint(175,255),random.randint(175,255),random.randint(175,255))
rcol = '#%02X%02X%02X' % (random.randint(175,255),random.randint(175,255),random.randint(175,255))
cols[chksum] = fcol,rcol
qual = {"note" : ["olp_{}".format(chksum)],
"chksum" : [chksum],
"ApEinfo_fwdcolor" : [fcol],
"ApEinfo_revcolor" : [rcol]}
if not chksum in [f.qualifiers["chksum"][0] for f in a.features if f.type == "overlap"]:
a.features.append( SeqFeature( FeatureLocation(start_in_a,
start_in_a + length),
type = "overlap",
qualifiers = qual))
if not chksum in [f.qualifiers["chksum"][0] for f in b.features if f.type == "overlap"]:
b.features.append( SeqFeature( FeatureLocation(start_in_b,
start_in_b + length),
type = "overlap",
qualifiers = qual))
for ds in dsset:
ds.features = sorted([f for f in ds.features], key = operator.attrgetter("location.start"))
self.analyzed_dsrecs = list(dsset)
# Create graph
self.G=nx.MultiDiGraph(multiedges=True, name ="original graph" , selfloops=False)
self.G.add_node( '5' )
self.G.add_node( '3' )
for i, dsrec in enumerate(self.analyzed_dsrecs):
overlaps = sorted( list({f.qualifiers['chksum'][0]:f for f in dsrec.features
if f.type=='overlap'}.values()),
key = operator.attrgetter('location.start'))
if overlaps:
overlaps = ([SeqFeature(FeatureLocation(0, 0),
type = 'overlap',
qualifiers = {'chksum':['5']})]+
overlaps+
[SeqFeature(FeatureLocation(len(dsrec),len(dsrec)),
type = 'overlap',
qualifiers = {'chksum':['3']})])
for olp1, olp2 in itertools.combinations(overlaps, 2):
n1 = olp1.qualifiers['chksum'][0]
n2 = olp2.qualifiers['chksum'][0]
if n1 == '5' and n2=='3':
continue
s1,e1,s2,e2 = (olp1.location.start.position,
olp1.location.end.position,
olp2.location.start.position,
olp2.location.end.position,)
source_fragment = Fragment(dsrec,s1,e1,s2,e2,i)
self.G.add_edge( n1, n2,
frag=source_fragment,
weight = s1-e1,
i = i)
#linear assembly
linear_products=defaultdict(list)
for path in all_simple_paths_edges(self.G, '5', '3', data=True, cutoff=self.max_nodes):
pred_frag = copy(list(path[0][2].values()).pop()['frag'])
source_fragments = [pred_frag, ]
if pred_frag.start2<pred_frag.end1:
result=pred_frag[pred_frag.start2+(pred_frag.end1-pred_frag.start2):pred_frag.end2]
else:
result=pred_frag[pred_frag.end1:pred_frag.end2]
for first_node, second_node, edgedict in path[1:]:
edgedict = list(edgedict.values()).pop()
f = copy(edgedict['frag'])
f.alignment = pred_frag.alignment + pred_frag.start2- f.start1
source_fragments.append(f)
if f.start2>f.end1:
result+=f[f.end1:f.end2]
else:
result+=f[f.start2+(f.end1-f.start2):f.end2]
pred_frag = f
add=True
for lp in linear_products[len(result)]:
if (str(result.seq).lower() == str(lp.seq).lower()
or
str(result.seq).lower() == str(lp.seq.reverse_complement()).lower()):
add=False
for dsrec in self.dsrecs:
if (str(result.seq).lower() == str(dsrec.seq).lower()
or
str(result.seq).lower() == str(dsrec.seq.reverse_complement()).lower()):
add=False
if add:
linear_products[len(result)].append(Contig( result, source_fragments))
self.linear_products = list(itertools.chain.from_iterable(linear_products[size] for size in sorted(linear_products, reverse=True)))
# circular assembly
self.cG = self.G.copy()
self.cG.remove_nodes_from(('5','3'))
#circular_products=defaultdict(list)
circular_products={}
for pth in all_circular_paths_edges(self.cG):
ns = min(enumerate(pth), key = lambda x:x[1][2]['i'])[0]
path = pth[ns:]+pth[:ns]
pred_frag = copy(path[0][2]['frag'])
source_fragments = [pred_frag, ]
if pred_frag.start2<pred_frag.end1:
result=pred_frag[pred_frag.start2+(pred_frag.end1-pred_frag.start2):pred_frag.end2]
else:
result=pred_frag[pred_frag.end1:pred_frag.end2]
result.seq = Dseq(str(result.seq))
for first_node, second_node, edgedict in path[1:]:
f = copy(edgedict['frag'])
f.alignment = pred_frag.alignment + pred_frag.start2- f.start1
source_fragments.append(f)
if f.start2>f.end1:
nxt = f[f.end1:f.end2]
else:
nxt =f[f.start2+(f.end1-f.start2):f.end2]
nxt.seq = Dseq(str(nxt.seq))
result+=nxt
pred_frag = f
#add=True
#for cp in circular_products[len(result)]:
# if (str(result.seq).lower() in str(cp.seq).lower()*2
# or
# str(result.seq).lower() == str(cp.seq.reverse_complement()).lower()*2):
# pass
# add=False
# print "##--"
#if add:
# circular_products[len(result)].append( Contig( Dseqrecord(result, circular=True), source_fragments))
r = Dseqrecord(result, circular=True)
circular_products[r.cseguid()] = Contig(r, source_fragments )
#self.circular_products = list(itertools.chain.from_iterable(circular_products[size] for size in sorted(circular_products, reverse=True)))
self.circular_products = sorted(list(circular_products.values()), key=len, reverse=True)