def test_rem_dash(self):
# ****--****-------**** 4M2D4M7D4M
# *******-------******* 7M7D7M
# has 4 dash columns and should become
# ****--****---**** 4M2D4M3D4M
# *******---******* 7M3D7M
for i in range(100):
dash_cols = random.randint(0, 10)
tStart = random.randint(0, 1000)
qStart = random.randint(0, 1000)
epo_pair = (
EPOitem._strfactory("homo_sapiens\t0\t1\t%d\t%d\t1\t%s" % (tStart, tStart+12-1, "4M2D4M%dD4M" % (dash_cols+3))),
EPOitem._strfactory("mus_musculus\t0\t1\t%d\t%d\t1\t%s" % (qStart, qStart+14-1, "7M%dD7M" % (dash_cols+3))))
chain = Chain._make_from_epo(epo_pair[0], epo_pair[1], {"chr1": 500}, {"chr1": 800})
ti = epo_pair[0].intervals(False)
qi = epo_pair[1].intervals(False)
assert ti[2][0] - ti[1][1] - dash_cols == chain[2][1]
assert qi[1][0] - qi[0][1] - dash_cols == chain[2][1]
# ----*****
# *-------*
# has 3 dash cols and should become
# *
# *
# with the qStart += 1 and tStart += 4
for i in range(100):
dash_cols = random.randint(0, 10)
tm = random.randint(6, 10)
qm = random.randint(1, 5)
tStart = random.randint(0, 1000)
qStart = random.randint(0, 1000)
epo_pair = (
EPOitem._strfactory("homo_sapiens\t0\t1\t%d\t%d\t1\t%s" % (tStart, tStart+tm-1, "%dD%dM" % (dash_cols+1, tm))),
EPOitem._strfactory("mus_musculus\t0\t1\t%d\t%d\t1\t%s" % (qStart, qStart+qm+1-1, "M%dD%dM" % (dash_cols+tm-qm, qm))))
chain = Chain._make_from_epo(epo_pair[0], epo_pair[1], {"chr1": 500}, {"chr1": 800})
if chain[1][-1] != qm:
pdb.set_trace()
assert chain[1][-1] == qm
# correct also for coordinate interpretation differences between UCSC and EPO
assert (qStart + 1) - 1 == chain[0].qStart, "%d != %d" % (qStart + 1, chain[0].qStart)
def convert_action(trg_comp, qr_comp, ts, qs, opt):
for i, (a,b) in enumerate(product(trg_comp, qr_comp)):
try:
ch, S, T, Q = Chain._make_from_epo(a, b, ts, qs)
if np.sum(S) == 0:
log.info("insignificant genomic alignment block %s ..." % ch.id)
continue
new_id = "%si%d" % (ch.id, i)
print >>opt.output, str(ch._replace(id=new_id))
map(lambda tup: opt.output.write("%d %d %d\n" % tup), izip(S,T,Q))
print >>opt.output, "%d\n" % S[-1]
except KeyError:
log.warning("skipping chromosome/contig (%s, %s)" % (a.chrom, b.chrom))
def convert_action(trg_comp, qr_comp, ts, qs, opt):
for i, (a,b) in enumerate(product(trg_comp, qr_comp)):
try:
ch, S, T, Q = Chain._make_from_epo(a, b, ts, qs)
if np.sum(S) == 0:
log.info("insignificant genomic alignment block %s ..." % ch.id)
continue
new_id = "%si%d" % (ch.id, i)
print(str(ch._replace(id=new_id)), file=opt.output)
map(lambda tup: opt.output.write("%d %d %d\n" % tup), zip(S, T, Q))
print("%d\n" % S[-1], file=opt.output)
except KeyError:
log.warning("skipping chromosome/contig (%s, %s)" % (a.chrom, b.chrom))
def test_make_chain(self):
def cch(cigar, s, e):
return cigar[s:e].find('-') == -1
for p in self.epo_records:
chain = Chain._make_from_epo(p[0], p[1], {"chr1": 500}, {"chr1": 800})
if not chain:
continue
ch, S, T, Q = chain
i = int(ch.id)
c1, c2 = cigar_pairs[i]
if p[0].strand == '-':
c1 = c1[::-1]
c2 = c2[::-1]
th = 0
for s, t, q in zip(S, T, Q):
if not (cch(c1, th, th+s) and cch(c2, th, th+s)):
pdb.set_trace()
assert cch(c1, th, th+s) and cch(c2, th, th+s), "%s and %s" % (c1[th:th+s], c2[th:th+s])
if t > q:
cch(c1, th+s, th+s+t) and c1[th+s:th+s+t] == '-'*t
else:
cch(c2, th+s, th+s+q) and c1[th+s:th+s+q] == '-'*q
th = th + s + max(t, q)