def do_test_coal_sim(self, stree, gene2species, n,
ntrees=10000, tabsize=30):
"""Perform a coal gene tree simulation test"""
tops = []
lookup = {}
util.tic("simulating %d trees" % ntrees)
for i in xrange(ntrees):
tree, recon = coal.sample_multicoal_tree(stree, n,
namefunc=lambda x: x)
tops.append(phylo.hash_tree(tree))
lookup[tops[-1]] = (tree, recon)
util.toc()
hist = histtab(tops)
probs = []
for row in hist:
tree, recon= lookup[row["item"]]
try:
#treelib.draw_tree_names(tree, maxlen=5)
treelib.remove_single_children(tree)
nodes = set(tree.postorder())
for node, snode in recon.items():
if node not in nodes:
del recon[node]
p = coal.prob_coal_recon_topology(tree, recon, stree, n)
except:
draw_tree(tree, maxlen=5, minlen=5)
raise
probs.append(exp(p))
return hist, probs
def _test_multicoal_tree(stree, n, nsamples):
"""test multicoal_tree"""
tops = {}
for i in xrange(nsamples):
tree, recon = coal.sample_multicoal_tree(stree, n,
namefunc=lambda x: x)
top = phylo.hash_tree(tree)
tops.setdefault(top, [0, tree, recon])[0] += 1
tab = Table(headers=["top", "simple_top", "percent", "prob"])
for top, (num, tree, recon) in tops.items():
tree2 = tree.copy()
treelib.remove_single_children(tree2)
print phylo.hash_tree(tree2)
print phylo.hash_tree(stree)
tab.add(top=top,
simple_top=phylo.hash_tree(tree2),
percent=num/float(nsamples),
prob=exp(coal.prob_multicoal_recon_topology(
tree, recon, stree, n)))
tab.sort(col="prob", reverse=True)
return tab, tops
def test_cdf_bmc(self):
# test cdf mrca BMC
stree = treelib.parse_newick(
"((A:1000, B:1000):500, (C:700, D:700):800);")
n = 1000
gene_counts = dict.fromkeys(stree.leaf_names(), 1)
T = 2000
p = exp(coal.cdf_mrca_bounded_multicoal(gene_counts, T, stree, n))
nsamples = 5000
c = 0
for i in xrange(nsamples):
tree, recon = coal.sample_multicoal_tree(stree, n)
if treelib.get_tree_timestamps(tree)[tree.root] < T:
c += 1
p2 = c / float(nsamples)
fequal(p, p2, .05)
def do_test_coal_sim(self,
stree,
gene2species,
n,
ntrees=10000,
tabsize=30):
"""Perform a coal gene tree simulation test"""
tops = []
lookup = {}
util.tic("simulating %d trees" % ntrees)
for i in xrange(ntrees):
tree, recon = coal.sample_multicoal_tree(stree,
n,
namefunc=lambda x: x)
tops.append(phylo.hash_tree(tree))
lookup[tops[-1]] = (tree, recon)
util.toc()
hist = histtab(tops)
probs = []
for row in hist:
tree, recon = lookup[row["item"]]
try:
#treelib.draw_tree_names(tree, maxlen=5)
treelib.remove_single_children(tree)
nodes = set(tree.postorder())
for node, snode in recon.items():
if node not in nodes:
del recon[node]
p = coal.prob_coal_recon_topology(tree, recon, stree, n)
except:
draw_tree(tree, maxlen=5, minlen=5)
raise
probs.append(exp(p))
return hist, probs