def check_bootstrap(self, filename, format, align_type="d"):
"""Check we can use fseqboot to pseudosample an alignment.
The align_type type argument is passed to the commandline object to
set the output format to use (from [D]na,[p]rotein and [r]na )
"""
self.assertTrue(os.path.isfile(filename), "Missing %s" % filename)
cline = FSeqBootCommandline(exes["fseqboot"],
sequence=filename,
outfile="test_file",
seqtype=align_type,
reps=2,
auto=True,
filter=True)
stdout, stderr = cline()
# the resultant file should have 2 alignments...
with open("test_file") as handle:
bs = list(AlignIO.parse(handle, format))
self.assertEqual(len(bs), 2)
# ..and each name in the original alignment...
with open(filename) as handle:
a_names = [
s.name.replace(" ", "_") for s in AlignIO.read(handle, format)
]
# ...should be in each alignment in the bootstrapped file
for a in bs:
self.assertEqual(a_names, [s.name.replace(" ", "_") for s in a])
def runSeqBoot(phylipfile, outbase):
"""
This part runs seqboot for the phylip file
"""
print "Running Seqboot"
seqout = outbase + ".fseqboot"
seqboot_cline = FSeqBootCommandline(sequence=phylipfile,
reps=1000,
seed=333,
outfile=seqout)
stdout, stderr = seqboot_cline()
return seqout
def build_nj_phylip(alignment, outfile, outgroup, work_dir="."):
"""
build neighbor joining tree of DNA seqs with PHYLIP in EMBOSS
PHYLIP manual
http://evolution.genetics.washington.edu/phylip/doc/
"""
phy_file = op.join(work_dir, "work", "aln.phy")
try:
AlignIO.write(alignment, file(phy_file, "w"), "phylip")
except ValueError:
print >>sys.stderr, \
"Repeated seq name, possibly due to truncation. NJ tree not built."
return None
seqboot_out = phy_file.rsplit(".", 1)[0] + ".fseqboot"
seqboot_cl = FSeqBootCommandline(FPHYLIP_BIN("fseqboot"), \
sequence=phy_file, outfile=seqboot_out, \
seqtype="d", reps=100, seed=12345)
stdout, stderr = seqboot_cl()
logging.debug("Resampling alignment: %s" % seqboot_cl)
dnadist_out = phy_file.rsplit(".", 1)[0] + ".fdnadist"
dnadist_cl = FDNADistCommandline(FPHYLIP_BIN("fdnadist"), \
sequence=seqboot_out, outfile=dnadist_out, method="f")
stdout, stderr = dnadist_cl()
logging.debug\
("Calculating distance for bootstrapped alignments: %s" % dnadist_cl)
neighbor_out = phy_file.rsplit(".", 1)[0] + ".njtree"
e = phy_file.rsplit(".", 1)[0] + ".fneighbor"
neighbor_cl = FNeighborCommandline(FPHYLIP_BIN("fneighbor"), \
datafile=dnadist_out, outfile=e, outtreefile=neighbor_out)
stdout, stderr = neighbor_cl()
logging.debug("Building Neighbor Joining tree: %s" % neighbor_cl)
consense_out = phy_file.rsplit(".", 1)[0] + ".consensustree.nodesupport"
e = phy_file.rsplit(".", 1)[0] + ".fconsense"
consense_cl = FConsenseCommandline(FPHYLIP_BIN("fconsense"), \
intreefile=neighbor_out, outfile=e, outtreefile=consense_out)
stdout, stderr = consense_cl()
logging.debug("Building consensus tree: %s" % consense_cl)
# distance without bootstrapping
dnadist_out0 = phy_file.rsplit(".", 1)[0] + ".fdnadist0"
dnadist_cl0 = FDNADistCommandline(FPHYLIP_BIN("fdnadist"), \
sequence=phy_file, outfile=dnadist_out0, method="f")
stdout, stderr = dnadist_cl0()
logging.debug\
("Calculating distance for original alignment: %s" % dnadist_cl0)
# infer branch length on consensus tree
consensustree1 = phy_file.rsplit(".", 1)[0] + ".consensustree.branchlength"
run_ffitch(distfile=dnadist_out0, outtreefile=consensustree1, \
intreefile=consense_out)
# write final tree
ct_s = Tree(consense_out)
if outgroup:
t1 = consensustree1 + ".rooted"
t2 = smart_reroot(consensustree1, outgroup, t1)
if t2 == t1:
outfile = outfile.replace(".unrooted", "")
ct_b = Tree(t2)
else:
ct_b = Tree(consensustree1)
nodesupport = {}
for node in ct_s.traverse("postorder"):
node_children = tuple(sorted([f.name for f in node]))
if len(node_children) > 1:
nodesupport[node_children] = node.dist / 100.
for k, v in nodesupport.items():
ct_b.get_common_ancestor(*k).support = v
print ct_b
ct_b.write(format=0, outfile=outfile)
try:
s = op.getsize(outfile)
except OSError:
s = 0
if s:
logging.debug("NJ tree printed to %s" % outfile)
return outfile, phy_file
else:
logging.debug("Something was wrong. NJ tree was not built.")
return None