def test_NexusComments(self):
"""Test the ability to parse nexus comments at internal and leaf nodes."""
# A tree with simple comments throughout the tree.
ts1b = "((12:0.13,19[&comment1]:0.13)[&comment2]:0.1,(20:0.171,11:0.171):0.13)[&comment3];"
tree = Trees.Tree(ts1b)
self.assertEqual(self._get_flat_nodes(tree),
[(None, 0.0, None, "[&comment3]"),
(None, 0.1, None, "[&comment2]"),
(None, 0.13, None, None), ("12", 0.13, None, None),
("19", 0.13, None, "[&comment1]"),
("20", 0.171, None, None),
("11", 0.171, None, None)])
# A tree with more complex comments throughout the tree.
# This is typical of the MCC trees produced by `treeannotator` in the beast-mcmc suite of phylogenetic tools
# The key difference being tested here is the ability to parse internal node comments that include ','.
ts1b = "(((9[&rate_range={1.3E-5,0.10958320752991428},height_95%_HPD={0.309132419999969,0.3091324199999691},length_range={3.513906814545109E-4,0.4381986285528381},height_median=0.309132419999969,length_95%_HPD={0.003011577063374571,0.08041621647998398}]:0.055354097721950546,5[&rate_range={1.3E-5,0.10958320752991428},height_95%_HPD={0.309132419999969,0.3091324199999691},length_range={3.865051168833178E-5,0.4391594442572986},height_median=0.309132419999969,length_95%_HPD={0.003011577063374571,0.08041621647998398}]:0.055354097721950546)[&height_95%_HPD={0.3110921040545068,0.38690865205576275},length_range={0.09675588357303178,0.4332959544380489},length_95%_HPD={0.16680375169879613,0.36500804261814374}]:0.20039426358269385)[&height_95%_HPD={0.5289500597932948,0.6973881165460601},length_range={0.02586430194846201,0.29509451958008265},length_95%_HPD={0.0840287249314221,0.2411078625957056}]:0.23042678598484334)[&height_95%_HPD={0.7527502510685965,0.821862094763501},height_median=0.8014438411766163,height=0.795965080422763,posterior=1.0,height_range={0.49863013698599995,0.821862094763501},length=0.0];"
tree = Trees.Tree(ts1b)
self.assertEqual(self._get_flat_nodes(tree), [
(None, 0.0, None,
"[&height_95%_HPD={0.7527502510685965,0.821862094763501},height_median=0.8014438411766163,height=0.795965080422763,posterior=1.0,height_range={0.49863013698599995,0.821862094763501},length=0.0]"
),
(None, 0.23042678598484334, None,
"[&height_95%_HPD={0.5289500597932948,0.6973881165460601},length_range={0.02586430194846201,0.29509451958008265},length_95%_HPD={0.0840287249314221,0.2411078625957056}]"
),
(None, 0.20039426358269385, None,
"[&height_95%_HPD={0.3110921040545068,0.38690865205576275},length_range={0.09675588357303178,0.4332959544380489},length_95%_HPD={0.16680375169879613,0.36500804261814374}]"
),
("9", 0.055354097721950546, None,
"[&rate_range={1.3E-5,0.10958320752991428},height_95%_HPD={0.309132419999969,0.3091324199999691},length_range={3.513906814545109E-4,0.4381986285528381},height_median=0.309132419999969,length_95%_HPD={0.003011577063374571,0.08041621647998398}]"
),
("5", 0.055354097721950546, None,
"[&rate_range={1.3E-5,0.10958320752991428},height_95%_HPD={0.309132419999969,0.3091324199999691},length_range={3.865051168833178E-5,0.4391594442572986},height_median=0.309132419999969,length_95%_HPD={0.003011577063374571,0.08041621647998398}]"
)
])
def test_internal_node_labels(self):
"""Handle text labels on internal nodes.
"""
ts1b = "(Cephalotaxus:125.000000,(Taxus:100.000000,Torreya:100.000000)"\
"TT1:25.000000)Taxaceae:90.000000;"
tree = Trees.Tree(ts1b)
self.assertEqual(self._get_flat_nodes(tree), [('Taxaceae', 90.0, None, None),
('Cephalotaxus', 125.0, None, None),
('TT1', 25.0, None, None),
('Taxus', 100.0, None, None),
('Torreya', 100.0, None, None)])
ts1c = "(Cephalotaxus:125.000000,(Taxus:100.000000,Torreya:100.000000)"\
"25.000000)90.000000;"
tree = Trees.Tree(ts1c)
self.assertEqual(self._get_flat_nodes(tree), [(None, 90.0, None, None),
('Cephalotaxus', 125.0, None, None),
(None, 25.0, None, None),
('Taxus', 100.0, None, None),
('Torreya', 100.0, None, None)])
ts2 = "(((t9:0.385832, (t8:0.445135,t4:0.41401)C:0.024032)B:0.041436,"\
"t6:0.392496)A:0.0291131, t2:0.497673, ((t0:0.301171,"\
"t7:0.482152)E:0.0268148, ((t5:0.0984167,t3:0.488578)G:0.0349662,"\
"t1:0.130208)F:0.0318288)D:0.0273876);"
tree = Trees.Tree(ts2)
large_ex_handle = open(os.path.join(self.testfile_dir,
"int_node_labels.nwk"))
tree = Trees.Tree(large_ex_handle.read())
large_ex_handle.close()
def test_merge_with_support(self):
"""Test merge_with_support and consensus method."""
ts1 = ("(((B 9:0.385832, (C 8:0.445135, C 4:0.41401)C:0.024032)B:0.041436,"
"A 6:0.392496)A:0.0291131, t2:0.497673, ((E 0:0.301171,"
"E 7:0.482152)E:0.0268148, ((G 5:0.0984167,G 3:0.488578)G:0.0349662,"
"F 1:0.130208)F:0.0318288)D:0.0273876);")
tbs1 = ("(((B 9:0.385832, (C 8:0.445135, C 4:0.41401)C:0.024032)B:0.041436,"
"A 6:0.392496)A:0.0291131, t2:0.497673, ((G 5:0.0984167,"
"G 3:0.488578)E:0.0268148, ((E 0:0.301171, E 7:0.482152)G:0.0349662,"
"F 1:0.130208)F:0.0318288)D:0.0273876);")
tbs2 = ("(((B 9:0.385832,A 6:0.392496 C:0.024032)B:0.041436, (C 8:0.445135,"
"C 4:0.41401))A:0.0291131, t2:0.497673, ((E 0:0.301171, E 7:0.482152)"
"E:0.0268148, ((G 5:0.0984167,G 3:0.488578)G:0.0349662,F 1:0.130208)"
"F:0.0318288)D:0.0273876);")
t1 = Trees.Tree(ts1)
tb1 = Trees.Tree(tbs1)
tb2 = Trees.Tree(tbs2)
t1.branchlength2support()
tb1.branchlength2support()
tb2.branchlength2support()
t1.merge_with_support(bstrees=[tb1, tb2], threshold=0.2)
supports = []
for i in t1.all_ids():
node = t1.node(i)
data = node.get_data()
supports.append(data.support)
self.assertTrue(supports, [0.0, 1.0, 0.04, 1.0, 0.5, 1.0, 1.0, 1.0,
1.0, 1.0, 0.5, 1.0, 1.0, 0.5, 1.0, 1.0, 1.0,
1.0])
def showConsensusTreeBio(self):
if self.path1 != '' and self.path2 != '':
# odczytaj rozszerzenie
self.fileEx1 = (os.path.splitext(self.path1)[1])[1:]
self.fileEx2 = (os.path.splitext(self.path2)[1])[1:]
self.trees = []
# pierwsze
self.f = open(self.path1, 'r')
self.tree1 = Trees.Tree(self.f.read())
self.trees.append(self.tree1)
self.f.close()
# drugie
self.f = open(self.path2, 'r')
self.tree2 = Trees.Tree(self.f.read())
self.trees.append(self.tree2)
self.f.close()
self.consensus_tree = Trees.consensus(self.trees)
# SHOWTIME
# rysuj
self.handle = StringIO(
self.consensus_tree.to_string(plain_newick=True))
self.tree = Phylo.read(self.handle, 'newick')
#self.tree.root.color = '#808080'
self.OpenInfoWindow()
else:
print "Nie wybrano punktow"
# WYSWIETLA INFORMACJE
img = pylab.imread('img/wally.png', 'rb')
pylab.imshow(img)
pylab.plot(0, 0)
# DAJE CZYSTY OBRAZ BEZ OSI ** PEWNIE MOZNA PROSCIEJ
frame1 = pylab.gca()
for xlabel_i in frame1.axes.get_xticklabels():
xlabel_i.set_visible(False)
xlabel_i.set_fontsize(0.0)
for xlabel_i in frame1.axes.get_yticklabels():
xlabel_i.set_fontsize(0.0)
xlabel_i.set_visible(False)
for tick in frame1.axes.get_xticklines():
tick.set_visible(False)
for tick in frame1.axes.get_yticklines():
tick.set_visible(False)
# SHOWTIME
pylab.show()
def drawConsensusTreeBioNexus(self):
if self.path1 != '' and self.path2 != '':
#get files extensions
self.fileExtension1 = (os.path.splitext(self.path1)[1])[1:]
self.fileExtension2 = (os.path.splitext(self.path2)[1])[1:]
#open tree files
self.trees = []
#first tree
self.f = open(self.path1, 'r')
self.tree1 = Trees.Tree(self.f.read())
self.trees.append(self.tree1)
self.f.close()
#second tree
self.f = open(self.path2, 'r')
self.tree2 = Trees.Tree(self.f.read())
self.trees.append(self.tree2)
self.f.close()
#generate consensus tree
self.consensus_tree = Trees.consensus(self.trees)
#draw tree
self.handle = StringIO(self.consensus_tree.to_string(plain_newick=True))
self.tree = Phylo.read(self.handle, 'newick')
self.tree.root.color = '#808080'
Phylo.draw(self.tree)
def makeRootUnroot(self, mod):
if self.path1 != '' and self.path2 == '':
# get files extensions
self.fileEx1 = (os.path.splitext(self.path1)[1])[1:]
self.fileEx2 = (os.path.splitext(self.path2)[1])[1:]
# open tree files
self.trees = []
self.drzewo = []
# first tree
self.f = open(self.path1, 'r')
self.miss = self.f.read()
self.tree1 = Trees.Tree(self.miss)
self.dre = ts.Tree(self.miss)
print "# Before modification"
print self.tree1
if mod == 0:
print "# After modification -- Rooting (at midpoint):"
self.dre.root_midpoint()
elif mod == 1:
print "# After modification -- UnRooting:"
self.dre.unroot()
elif mod == 2:
print "# After modification -- Rooting (balanced):"
self.dre.root_balanced()
print self.dre
print "\nDetails about tree:"
self.dre.display()
Phylo.draw_ascii(self.tree1)
self.show()
self.f.close()
def parse_trees(filename):
"""Parse trees.
Helper function until we have Bio.Phylo on trunk.
"""
# TODO - Can this be removed now?
data = open("test_file", "r").read()
for tree_str in data.split(";\n"):
if tree_str:
yield Trees.Tree(tree_str + ";")
def test_to_string(self):
"""Test to_string method."""
t = Trees.Tree(
"(((B 9:0.385832, (C 8:0.445135, C 4:0.41401)C:0.024032)B:0.041436,"
"A 6:0.392496)A:0.0291131, t2:0.497673, ((E 0:0.301171,"
"E 7:0.482152)E:0.0268148, ((G 5:0.0984167,G 3:0.488578)G:0.0349662,"
"F 1:0.130208)F:0.0318288)D:0.0273876);")
tree_a_val = "((A 6,(B 9,(C 8,C 4))),t2,((E 0,E 7),(F 1,(G 5,G 3))));"
self.assertEqual(t.to_string(ladderize="LEFT"),
"tree a_tree = " + tree_a_val)
def showConsensusTreeBio(self):
if self.path1 != '' and self.path2 != '':
# odczytaj rozszerzenie
self.fileEx1 = (os.path.splitext(self.path1)[1])[1:]
self.fileEx2 = (os.path.splitext(self.path2)[1])[1:]
self.trees = []
# pierwsze
self.f = open(self.path1, 'r')
self.tree1 = Trees.Tree(self.f.read())
self.trees.append(self.tree1)
self.f.close()
# drugie
self.f = open(self.path2, 'r')
self.tree2 = Trees.Tree(self.f.read())
self.trees.append(self.tree2)
self.f.close()
self.consensus_tree = Trees.consensus(self.trees)
# SHOWTIME
# rysuj
self.handle = StringIO(self.consensus_tree.to_string(plain_newick=True))
self.tree = Phylo.read(self.handle, 'newick')
#self.tree.root.color = '#808080'
self.OpenInfoWindow()
else:
print "Nie wybrano punktow"
# WYSWIETLA INFORMACJE
img = pylab.imread('img/wally.png', 'rb')
pylab.imshow(img)
pylab.plot(0, 0)
# DAJE CZYSTY OBRAZ BEZ OSI ** PEWNIE MOZNA PROSCIEJ
frame1 = pylab.gca()
for xlabel_i in frame1.axes.get_xticklabels():
xlabel_i.set_visible(False)
xlabel_i.set_fontsize(0.0)
for xlabel_i in frame1.axes.get_yticklabels():
xlabel_i.set_fontsize(0.0)
xlabel_i.set_visible(False)
for tick in frame1.axes.get_xticklines():
tick.set_visible(False)
for tick in frame1.axes.get_yticklines():
tick.set_visible(False)
# SHOWTIME
pylab.show()
def showOpenFileWindow(self):
fname = QtGui.QFileDialog.getOpenFileName(self, 'Otworz plik zawierajacy drzewo', './Trees')
if fname != '':
self.chosenFileName = str(fname)
f = open(str(fname), 'r')
self.tree1 = Trees.Tree(f.read())
self.moredata = Trees.Tree.display(self.tree1)
print self.moredata
self.tmpf = open('more.txt', 'w')
self.tmpf.write(str(self.moredata))
self.tmpf.close()
# with f:
# path1 = f.read()
# self.pathInfo.setText(path1)
self.fileExtension = (os.path.splitext(str(fname))[1])[1:]
self.tree = Phylo.read(str(fname), self.fileExtension)
self.tree1s = self.tree
self.terminals = self.tree1s.get_terminals()
def test_large_newick(self):
with open(os.path.join(self.testfile_dir,
"int_node_labels.nwk")) as large_ex_handle:
tree = Trees.Tree(large_ex_handle.read())
def parse_trees(filename):
"""Helper function until we have Bio.Phylo on trunk."""
data = open("test_file", "r").read()
for tree_str in data.split(";\n"):
if tree_str:
yield Trees.Tree(tree_str + ";")
def test_TreeTest2(self):
"""Handle text labels on internal nodes."""
ts1b = (
"(Cephalotaxus:125.000000,(Taxus:100.000000,Torreya:100.000000)"
"TT1:25.000000)Taxaceae:90.000000;")
tree = Trees.Tree(ts1b)
self.assertEqual(self._get_flat_nodes(tree),
[("Taxaceae", 90.0, None, None),
("Cephalotaxus", 125.0, None, None),
("TT1", 25.0, None, None),
("Taxus", 100.0, None, None),
("Torreya", 100.0, None, None)])
tree.prune("Torreya")
self.assertEqual(tree.all_ids(), [0, 1, 3])
ts1c = (
"(Cephalotaxus:125.000000,(Taxus:100.000000,Torreya:100.000000)"
"25.000000)90.000000;")
tree = Trees.Tree(ts1c)
self.assertEqual(self._get_flat_nodes(tree),
[(None, 90.0, None, None),
("Cephalotaxus", 125.0, None, None),
(None, 25.0, None, None),
("Taxus", 100.0, None, None),
("Torreya", 100.0, None, None)])
self.assertFalse(tree.has_support())
with self.assertRaises(Exception) as context:
tree.randomize()
self.assertTrue(
"Either numer of taxa or list of taxa must be specified." in str(
context.exception))
tree_rand = Trees.Tree(ts1c)
tree_rand.randomize(ntax=4)
self.assertEqual(sorted(tree_rand.get_taxa()),
["taxon1", "taxon2", "taxon3", "taxon4"])
tree.branchlength2support()
tree.convert_absolute_support(2)
self.assertEqual(self._get_flat_nodes(tree),
[(None, 0.0, 90.0, None),
("Cephalotaxus", 0.0, 62.5, None),
(None, 0.0, 12.5, None), ("Taxus", 0.0, 50.0, None),
("Torreya", 0.0, 50.0, None)])
ts2 = (
"(((t9:0.385832, (t8:0.445135,t4:0.41401)C:0.024032)B:0.041436,"
"t6:0.392496)A:0.0291131, t2:0.497673, ((t0:0.301171,"
"t7:0.482152)E:0.0268148, ((t5:0.0984167,t3:0.488578)G:0.0349662,"
"t1:0.130208)F:0.0318288)D:0.0273876);")
tree = Trees.Tree(ts2)
tree.branchlength2support()
supports = []
for i in tree.all_ids():
node = tree.node(i)
data = node.get_data()
supports.append(data.support)
self.assertEqual(supports, [
0.0, 0.0291131, 0.041436, 0.385832, 0.024032, 0.445135, 0.41401,
0.392496, 0.497673, 0.0273876, 0.0268148, 0.301171, 0.482152,
0.0318288, 0.0349662, 0.0984167, 0.488578, 0.130208
])
ts3 = (
"(((B 9:0.385832, (C 8:0.445135, C4:0.41401)C:0.024032)B:0.041436,"
"A 6:0.392496)A:0.0291131, t2:0.497673, ((E 0:0.301171,"
"E 7:0.482152)E:0.0268148, ((G 5:0.0984167,G 3:0.488578)G:0.0349662,"
"F 1:0.130208)F:0.0318288)D:0.0273876);")
self.assertFalse(tree.is_identical(Trees.Tree(ts3)))
tree = Trees.Tree(ts3)
self.assertTrue(tree.is_bifurcating())
self.assertTrue(tree.is_bifurcating(1))
self.assertEqual([tree.distance(0, n) for n in tree.all_ids()], [
0.0, 0.0291131, 0.0705491, 0.4563811, 0.0945811, 0.5397161,
0.5085911, 0.4216091, 0.497673, 0.0273876, 0.0542024, 0.3553734,
0.5363544, 0.0592164, 0.09418259999999999, 0.1925993, 0.5827606,
0.1894244
])
subtree = tree.set_subtree(10)
self.assertEqual(sorted(subtree), ["E 0", "E 7"])
tree.collapse_genera()
self.assertEqual(tree.all_ids(),
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 13, 14, 17])
if options.tree=="":
DoError("No tree file specified")
print "Reading tree file"
sys.stdout.flush()
try:
tree_strings = open(options.tree).readlines()
except IOError:
DoError("Cannot open tree file "+options.tree)
handle = open(options.outputfile, "w")
for tree_string in tree_strings:
tree = Trees.Tree(tree_string, rooted=True)
if options.outgroup!="":
print "Rooting tree on", options.outgroup
sys.stdout.flush()
tree.root_with_outgroup(outgroup=options.outgroup)
else:
print "Midpoint rooting tree"
tree=Si_nexus.midpoint_root(tree)
#Print tree
treestring=Si_nexus.tree_to_string(tree,plain=False,ladderize=options.ladderise)#tree with branch lengths and support
def makeJsonAlign(smo_output, phy_output):
_writer = json.JsonWriter()
#Making JSON for root alignment in .smo file
treeStr = readTree(phy_output)
#Find left and right IDs using MPTT
tree = Trees.Tree(rooted=True)
(clades, vals) = tree._parse(treeStr)
tree._add_subtree(0, clades)
rootNode = tree.node(tree.root)
nodeIdOfLeftId = {}
leftIdOfSatchmoId = {}
mptt(tree, 1, 0, nodeIdOfLeftId)
#tree.display()
#print(nodeIdOfLeftId)
smoFile = smo_output
#Read from .smo file
f_smo = open(smoFile)
smoString = ''.join([line for line in f_smo.readlines()])
f_smo.close()
#Find the number of nodes
index1 = string.find(smoString, ' ')
index2 = string.find(smoString, '\n')
no_of_nodes = int(smoString[index1 + 1:index2])
#Read sequence alignment at each node
curr_index = 0
i = 0
ret = {}
keys = {}
no_of_leaves = {}
while i < no_of_nodes:
curr_index = string.find(smoString, '[', curr_index)
leaf_index = string.find(smoString, "leaf", curr_index)
internal_index = string.find(smoString, "internal", curr_index)
if ((leaf_index < internal_index) or
(internal_index == -1)) and (leaf_index != -1):
nodeType = "leaf"
#print("Its a leaf!")
elif ((leaf_index > internal_index) or
(leaf_index == -1)) and (internal_index != -1):
nodeType = "internal"
#print("Its an internal node!")
else:
print("Invalid node type!")
if (nodeType == "leaf"):
curr_index = string.find(smoString, "index", curr_index)
newLine_index = string.find(smoString, '\n', curr_index)
satchmo_id = int(smoString[curr_index + len("index") +
1:newLine_index])
curr_index = string.find(smoString, '>', curr_index)
newLine_index = string.find(smoString, '\n', curr_index)
seq_name = smoString[curr_index + 1:newLine_index]
for id in nodeIdOfLeftId.keys():
if (tree.is_terminal(
nodeIdOfLeftId[id])) and (seq_name in tree.get_taxa(
nodeIdOfLeftId[id])):
leftIdOfSatchmoId[satchmo_id] = id
populate_ret(ret, id, curr_index, keys, no_of_leaves,
smoString)
elif (nodeType == "internal"):
curr_index = string.find(smoString, "index", curr_index)
newLine_index = string.find(smoString, '\n', curr_index)
satchmo_id = int(smoString[curr_index + len("index") +
1:newLine_index])
target_index = string.find(smoString, "target", curr_index)
newLine_index = string.find(smoString, '\n', target_index)
target = int(smoString[target_index + len("target") +
1:newLine_index])
target = nodeIdOfLeftId[leftIdOfSatchmoId[target]]
#print("Target = ." + str(target) + ".");
template_index = string.find(smoString, "template", curr_index)
newLine_index = string.find(smoString, '\n', template_index)
template = int(smoString[template_index + len("template") +
1:newLine_index])
template = nodeIdOfLeftId[leftIdOfSatchmoId[template]]
#print("Template = ." + str(template) + ".")
#print(tree.node(nodeIdOfLeftId[id]).get_succ())
for id in nodeIdOfLeftId.keys():
#print(tree.node(nodeIdOfLeftId[id]).get_succ())
if(not(tree.is_terminal(nodeIdOfLeftId[id]))) \
and (template in tree.node(nodeIdOfLeftId[id]).get_succ()) \
and (target in tree.node(nodeIdOfLeftId[id]).get_succ()) \
and (len(tree.node(nodeIdOfLeftId[id]).get_succ()) == 2):
#print("Template = ." + str(template) + ".")
#print(tree.node(nodeIdOfLeftId[id]).get_succ())
#print("Adding to id = " + str(id))
leftIdOfSatchmoId[satchmo_id] = id
curr_index = string.find(smoString, '>', curr_index)
populate_ret(ret, id, curr_index, keys, no_of_leaves,
smoString)
i += 1
#print(leftIdOfSatchmoId)
#print(ret)
for id in nodeIdOfLeftId.keys():
addColorString(ret, keys, no_of_leaves, id)
#print(ret)
ans = _writer.write(ret)
return ans