def test_prob_coal2(self):
outdir = 'test/tmp/test_coal/Coal_test_prob_coal2/'
make_clean_dir(outdir)
k = 2
n = 1000
p = Gnuplot()
p.enableOutput(False)
p.plotfunc(lambda t: coal.prob_coal(t, k, n), 0, 4000, 10,
ymin=0)
# draw single coal samples
x = [coal.sample_coal(k, n) for i in xrange(200)]
plotdistrib(x, 40, plot=p)
p.enableOutput(True)
p.save(outdir + 'plot.png')
eq_sample_pdf(x, lambda t: coal.prob_coal(t, k, n), 40)
def test_prob_coal_cond_counts_simple(self):
# when we condition on b=1, it is the same as the bounded coal
# PDF.
# prob_coal_cond_counts is actually a more general version of
# prob_bounded_coal
outdir = 'test/tmp/test_coal/Coal_test_prob_coal_counys_simple/'
make_clean_dir(outdir)
a = 5
b = 1
t = 2000
n = 1000
p = Gnuplot()
p.enableOutput(False)
for x in frange(0, 2000, 10):
y = coal.prob_coal_cond_counts_simple(x, a, b, t, n)
y2 = coal.prob_bounded_coal(x, a, n, t)
self.assertAlmostEqual(y, y2)
def test_cdf_mrca(self):
outdir = 'test/tmp/test_coal/Coal_test_cdf_mrca/'
make_clean_dir(outdir)
n = 1000
k = 6
step = 10
x = list(frange(0, 5000, step))
y = [coal.prob_mrca(i, k, n) * step for i in x]
y2 = cumsum(y)
y3 = [coal.cdf_mrca(t, k, n) for t in x]
p = Gnuplot()
p.enableOutput(False)
p.plot(x, y2, style="lines")
p.plot(x, y3, style="lines")
p.enableOutput(True)
p.save(outdir + 'plot.png')
eq_sample_pdf(x, lambda t: coal.cdf_mrca(t, k, n), 40)
def test_plot_prob_bounded_coal(self):
n = 1000
k = 4
t = 800
alltimes = []
# sample times
for i in xrange(5000):
while True:
times = [0]
for j in xrange(k, 1, -1):
times.append(times[-1] + coal.sample_coal(j, n))
if times[-1] >= t:
break
if times[-1] < t:
break
alltimes.append(times)
p = Gnuplot()
for i in range(1, 2):
x, y = distrib([q[i] - q[i-1] for q in alltimes], width=20)
p.plot(x, y, style="lines", xmax=500)
x = list(frange(0, 500, 10))
#for i in range(1, 2): #k):
y2 = [coal.prob_bounded_coal(j, k, n, t) for j in x]
p.plot(x, y2, style="lines", xmax=500)
fequals(y, y2, rel=.05, eabs=.01)
def test_coin_sample_post(self):
"""Test sampling from posterior distribution"""
outdir = 'test/tmp/test_hmm/test_coin_sample_post/'
make_clean_dir(outdir)
model = make_coin_model()
# sample states and data
ndata = 100
states = list(islice(hmm.sample_hmm_states(model), ndata))
data = list(hmm.sample_hmm_data(model, states))
model.prob_emission = (
lambda pos, state: model.prob_emission_data(state, data[pos]))
p = Gnuplot()
p.enableOutput(False)
p.plot(states, style="lines")
probs = hmm.get_posterior_probs(model, len(data))
states2 = [exp(probs[i][1]) for i in xrange(len(data))]
p.plot(util.vadds(states2, 1.5), style="lines", miny=-1, maxy=12)
for i in range(2, 10):
states2 = hmm.sample_posterior(model, ndata)
self.assertTrue(stats.corr(states, states2) > .5)
p.plot(util.vadds(states2, 1.5 * i),
style="lines",
miny=-1,
maxy=12)
p.enableOutput(True)
p.save(outdir + 'plot.png')
def test_coin(self):
"""Test that viterbi and posterior coding work well."""
outdir = 'test/tmp/test_hmm/test_coin/'
make_clean_dir(outdir)
model = make_coin_model()
# sample states
ndata = 100
states = list(islice(hmm.sample_hmm_states(model), ndata))
p = Gnuplot()
p.enableOutput(False)
p.plot(states, style="lines")
# sample data
data = list(hmm.sample_hmm_data(model, states))
# viterbi
model.prob_emission = (
lambda pos, state: model.prob_emission_data(state, data[pos]))
states2 = hmm.viterbi(model, len(data))
# posterior
probs = hmm.get_posterior_probs(model, len(data))
states3 = [exp(probs[i][1]) for i in xrange(len(data))]
# assert that inferences correlates with true state
self.assertTrue(stats.corr(states, states2) > .5)
self.assertTrue(stats.corr(states, states3) > .5)
# plot inference
p.plot(util.vadds(states2, 1.5), style="lines", miny=-1, maxy=4)
p.plot(util.vadds(states3, 2.5), style="lines", miny=-1, maxy=4)
p.enableOutput(True)
p.save(outdir + 'plot.png')
def test_coin_sample_post(self):
"""Test sampling from posterior distribution"""
outdir = 'test/tmp/test_hmm/test_coin_sample_post/'
make_clean_dir(outdir)
model = make_coin_model()
# sample states and data
ndata = 100
states = list(islice(hmm.sample_hmm_states(model), ndata))
data = list(hmm.sample_hmm_data(model, states))
model.prob_emission = (lambda pos, state:
model.prob_emission_data(state, data[pos]))
p = Gnuplot()
p.enableOutput(False)
p.plot(states, style="lines")
probs = hmm.get_posterior_probs(model, len(data))
states2 = [exp(probs[i][1]) for i in xrange(len(data))]
p.plot(util.vadds(states2, 1.5), style="lines", miny=-1, maxy=12)
for i in range(2, 10):
states2 = hmm.sample_posterior(model, ndata)
self.assertTrue(stats.corr(states, states2) > .5)
p.plot(util.vadds(states2, 1.5*i), style="lines", miny=-1, maxy=12)
p.enableOutput(True)
p.save(outdir + 'plot.png')
def test_coin(self):
"""Test that viterbi and posterior coding work well."""
outdir = 'test/tmp/test_hmm/test_coin/'
make_clean_dir(outdir)
model = make_coin_model()
# sample states
ndata = 100
states = list(islice(hmm.sample_hmm_states(model), ndata))
p = Gnuplot()
p.enableOutput(False)
p.plot(states, style="lines")
# sample data
data = list(hmm.sample_hmm_data(model, states))
# viterbi
model.prob_emission = (lambda pos, state:
model.prob_emission_data(state, data[pos]))
states2 = hmm.viterbi(model, len(data))
# posterior
probs = hmm.get_posterior_probs(model, len(data))
states3 = [exp(probs[i][1]) for i in xrange(len(data))]
# assert that inferences correlates with true state
self.assertTrue(stats.corr(states, states2) > .5)
self.assertTrue(stats.corr(states, states3) > .5)
# plot inference
p.plot(util.vadds(states2, 1.5), style="lines", miny=-1, maxy=4)
p.plot(util.vadds(states3, 2.5), style="lines", miny=-1, maxy=4)
p.enableOutput(True)
p.save(outdir + 'plot.png')
def test_top(self):
outdir = 'test/tmp/test_coal/BMC_test_top/'
make_clean_dir(outdir)
stree = treelib.parse_newick(
"(((A:200, E:200):800, B:1000):500, (C:700, D:700):800);")
n = 500
T = 2000
nsamples = 4000
# compare top hist with simpler rejection sampling
tops = {}
tops2 = {}
for i in xrange(nsamples):
# use rejection sampling
tree, recon = coal.sample_bounded_multicoal_tree_reject(
stree, n, T, namefunc=lambda x: x)
# sample tree
tree2, recon2 = coal.sample_bounded_multicoal_tree(
stree, n, T, namefunc=lambda x: x)
top = phylo.hash_tree(tree)
top2 = phylo.hash_tree(tree2)
tops.setdefault(top, [0, tree, recon])[0] += 1
tops.setdefault(top2, [0, tree2, recon2])
tops2.setdefault(top2, [0, tree2, recon2])[0] += 1
tops2.setdefault(top, [0, tree, recon])
keys = tops.keys()
x = [safelog(tops[i][0], default=0) for i in keys]
y = [safelog(tops2[i][0], default=0) for i in keys]
self.assertTrue(stats.corr(x, y) > .9)
p = Gnuplot()
p.enableOutput(False)
p.plot(x, y)
p.plot([min(x), max(x)], [min(x), max(x)], style="lines")
p.enableOutput(True)
p.save(outdir + 'plot.png')
def test_fast_sample_bounded_coal(self):
# sample bounded coal times efficiently
n = 1000
k = 5
t = 500
alltimes = []
# sample times
for i in xrange(5000):
while True:
times = [0]
for j in xrange(k, 1, -1):
times.append(times[-1] + coal.sample_coal(j, n))
if times[-1] >= t:
break
if times[-1] < t:
break
alltimes.append(times)
p = Gnuplot()
for i in range(1, k):
x, y = distrib([q[i] - q[i-1] for q in alltimes], width=30)
p.plot(x, y, style="lines", xmax=500)
p.enableOutput(True)
p.replot()
# sample times efficently
alltimes2 = []
for i in xrange(5000):
times = [0]
for j in xrange(k, 1, -1):
times.append(times[-1] +
coal.sample_bounded_coal(j, n, t-times[-1]))
alltimes2.append(times)
#p = Gnuplot()
for i in range(1, k):
x, y = distrib([q[i] - q[i-1] for q in alltimes2], width=30)
p.plot(x, y, style="lines", xmax=500)
p.enableOutput(True)
p.replot()
def test_cdf_coal_cond_counts(self):
# test coalescent pdf when conditioned on future lineage counts
outdir = 'test/tmp/test_coal/Coal_test_cdf_coal_cond_counts/'
make_clean_dir(outdir)
a = 5
for b in xrange(2, a):
t = 500
n = 1000
p = Gnuplot()
p.enableOutput(False)
p.plotfunc(lambda x: coal.cdf_coal_cond_counts(
x, a, b, t, n), 0, t, 10)
# draw single coal samples using rejection sampling
s = []
for i in xrange(1000):
while True:
times = coal.sample_coal_times(a, n)
if times[a-b-1] < t and (b == 1 or times[a-b] > t):
break
s.append(times[0])
x2, y2 = stats.cdf(s)
p.plot(x2, y2, style='lines')
p.enableOutput(True)
p.save(outdir + 'plot-%d.png' % b)
eq_sample_pdf(
x2, lambda x: coal.prob_coal_cond_counts(x, a, b, t, n), 40)
