def test_mk_piNotSumTo1_raisesAssertionError(self):
tree = self.randTree100Scale2
chars = self.simChars100states3Scale2
Q = np.array([[-0.556216, 0.278108, 0.278108],
[0.278108, -0.556216, 0.278108],
[0.278108, 0.278108, -0.556216]])
pi = np.array([0.0, 0.0, 0.0])
try:
discrete.mk(tree, chars, Q, pi=pi)
self.fail("Assertion error not raised")
except AssertionError, e:
self.assertEquals("values of given pi must sum to 1", e.message)
def test_mk_piWrongType_raisesAssertionError(self):
tree = self.randTree100Scale2
chars = self.simChars100states3Scale2
Q = np.array([[-0.556216, 0.278108, 0.278108],
[0.278108, -0.556216, 0.278108],
[0.278108, 0.278108, -0.556216]])
pi = [0.0, 0.0, 1.0]
try:
discrete.mk(tree, chars, Q, pi=pi)
self.fail("Assertion error not raised")
except AssertionError, e:
self.assertEquals("pi must be str or numpy array", e.message)
def test_mk_wronglengthpi_raisesAssertionError(self):
tree = self.randTree100Scale2
chars = self.simChars100states3Scale2
Q = np.array([[-0.556216, 0.278108, 0.278108],
[0.278108, -0.556216, 0.278108],
[0.278108, 0.278108, -0.556216]])
pi = np.array([0.0, 0.0, 0.0, 1.0])
try:
discrete.mk(tree, chars, Q, pi=pi)
self.fail("Assertion error not raised")
except AssertionError, e:
self.assertEquals("length of given pi does not match Q dimensions",
e.message)
def test_mk_piWrongType_raisesAssertionError(self):
tree = self.randTree100Scale2
chars = self.simChars100states3Scale2
Q = np.array([[-0.556216,0.278108,0.278108],
[0.278108,-0.556216,0.278108],
[0.278108,0.278108,-0.556216]])
pi = [0.0,0.0,1.0]
try:
discrete.mk(tree, chars, Q, pi=pi)
self.fail("Assertion error not raised")
except AssertionError, e:
self.assertEquals("pi must be str or numpy array",
e.message)
def test_mk_piNotSumTo1_raisesAssertionError(self):
tree = self.randTree100Scale2
chars = self.simChars100states3Scale2
Q = np.array([[-0.556216,0.278108,0.278108],
[0.278108,-0.556216,0.278108],
[0.278108,0.278108,-0.556216]])
pi = np.array([0.0,0.0,0.0])
try:
discrete.mk(tree, chars, Q, pi=pi)
self.fail("Assertion error not raised")
except AssertionError, e:
self.assertEquals("values of given pi must sum to 1",
e.message)
def test_mk_wronglengthpi_raisesAssertionError(self):
tree = self.randTree100Scale2
chars = self.simChars100states3Scale2
Q = np.array([[-0.556216,0.278108,0.278108],
[0.278108,-0.556216,0.278108],
[0.278108,0.278108,-0.556216]])
pi = np.array([0.0,0.0,0.0,1.0])
try:
discrete.mk(tree, chars, Q, pi=pi)
self.fail("Assertion error not raised")
except AssertionError, e:
self.assertEquals("length of given pi does not match Q dimensions",
e.message)
def test_mkFlatroot_randtree10_matchesPhytools(self):
charstates = self.randchars10
tree = self.randTree10
Q = self.randQ
phytoolslogLikelihood = -8.298437
calculatedLogLikelihood = discrete.mk(tree, charstates, Q)
self.assertTrue(np.isclose(phytoolslogLikelihood, calculatedLogLikelihood))
def test_mkFlatroot_randtree5_matchesPhytools(self):
charstates = self.randchars5
tree = self.randTree5
Q = self.randQ
phytoolslogLikelihood = -6.223166
calculatedLogLikelihood = discrete.mk(tree, charstates, Q)
self.assertTrue(
np.isclose(phytoolslogLikelihood, calculatedLogLikelihood))
def test_mkFlatroot_randtree10_matchesPhytools(self):
charstates = self.randchars10
tree = self.randTree10
Q = self.randQ
phytoolslogLikelihood = -8.298437
calculatedLogLikelihood = discrete.mk(tree, charstates, Q)
self.assertTrue(
np.isclose(phytoolslogLikelihood, calculatedLogLikelihood))
def test_mkFlatroot_randtree5_matchesPhytools(self):
charstates = self.randchars5
tree = self.randTree5
Q = self.randQ
phytoolslogLikelihood = -6.223166
calculatedLogLikelihood = discrete.mk(tree, charstates, Q)
self.assertTrue(np.isclose(phytoolslogLikelihood, calculatedLogLikelihood))
def test_mk_fitzjohn_matchesDiversitree(self):
charstates = [0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 2, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 1]
tree = ivy.tree.read("support/randtree100tipsscale2.newick")
Q = np.array([[-2.09613850e-01, 1.204029e-01, 8.921095e-02],
[5.654382e-01, -5.65438217e-01, 1.713339e-08],
[2.415020e-06, 5.958744e-07, -3.01089440e-06]])
expectedLikelihood = -32.79025
calculatedLogLikelihood = discrete.mk(tree, charstates, Q,
pi ="Fitzjohn")
self.assertTrue(np.isclose(expectedLikelihood, calculatedLogLikelihood))
def test_mkMultiRegime_sameQ_matchesmk(self):
tree = self.randTree100Scale2
chars = self.simChars100states3Scale2
Q = np.array([[-0.556216,0.278108,0.278108],
[0.278108,-0.556216,0.278108],
[0.278108,0.278108,-0.556216]])
Qs = np.array([Q,Q])
locs = [[i for i,n in enumerate(tree.postiter()) if not n.isroot][:50],
[i for i,n in enumerate(tree.postiter()) if not n.isroot][50:]]
single = discrete.mk(tree, chars, Q)
multi = discrete.mk_multi_regime(tree, chars, Qs, locs, pi="Equal")
self.assertEquals(single, multi)
def test_mkMultiRegime_sameQ_matchesmk(self):
tree = self.randTree100Scale2
chars = self.simChars100states3Scale2
Q = np.array([[-0.556216, 0.278108, 0.278108],
[0.278108, -0.556216, 0.278108],
[0.278108, 0.278108, -0.556216]])
Qs = np.array([Q, Q])
locs = [[i for i, n in enumerate(tree.postiter())
if not n.isroot][:50],
[i for i, n in enumerate(tree.postiter())
if not n.isroot][50:]]
single = discrete.mk(tree, chars, Q)
multi = discrete.mk_multi_regime(tree, chars, Qs, locs, pi="Equal")
self.assertEquals(single, multi)
def test_mkFlatroot_3tiptreeSymmetricQ2x2_returnslikelihood(self):
tree = self.threetiptree
chars = self.charstates_011
Q = self.Q2x2_sym
# Manually calculated likelihood for expected output
L0A = 1
L1A = 0
L0B = 0
L1B = 1
L0D = 0
L1D = 1
P00A = 0.90936538
P01A = 0.09063462
P11A = 0.90936538
P10A = 0.09063462
P00B = 0.90936538
P01B = 0.09063462
P11B = 0.90936538
P10B = 0.09063462
P00C = 0.90936538
P01C = 0.09063462
P11C = 0.90936538
P10C = 0.09063462
P00D = 0.83516002
P01D = 0.16483998
P11D = 0.83516002
P10D = 0.16483998
L0C = (P00A * L0A + P01A * L1A) * (P00B * L0B + P01B * L1B)
L1C = (P10A * L0A + P11A * L1A) * (P10B * L0B + P11B * L1B)
L0r = (P00C * L0C + P01C * L1C) * (P00D * L0D + P01D * L1D)
L1r = (P10C * L0C + P11C * L1C) * (P10D * L0D + P11D * L1D)
predictedLikelihood = math.log(L0r * 0.5 + L1r * 0.5)
calculatedLikelihood = discrete.mk(tree, chars, Q)
self.assertTrue(np.isclose(predictedLikelihood, calculatedLikelihood))
def test_mkFlatroot_2tiptreeSymmetricQ2x2difblens_returnslikelihood(self):
charstates = self.charstates_01
tree = self.simpletreedifblens
for i, node in enumerate(tree.leaves()):
node.charstate = charstates[i]
Q = self.Q2x2_sym
t = [ node.length for node in tree.descendants() ]
t = np.array(t, dtype=np.double)
p = cyexpokit.dexpm_tree(Q,t)
for i, node in enumerate(tree.descendants()):
node.pmat = p[i]
node = tree
predictedLikelihood = math.log(0.11279709097648889)
calculatedLikelihood = discrete.mk(tree, charstates, Q)
self.assertTrue(np.isclose(predictedLikelihood, calculatedLikelihood))
def test_mkFlatroot_2tiptreeSymmetricQ2x2difblens_returnslikelihood(self):
charstates = self.charstates_01
tree = self.simpletreedifblens
for i, node in enumerate(tree.leaves()):
node.charstate = charstates[i]
Q = self.Q2x2_sym
t = [node.length for node in tree.descendants()]
t = np.array(t, dtype=np.double)
p = cyexpokit.dexpm_tree(Q, t)
for i, node in enumerate(tree.descendants()):
node.pmat = p[i]
node = tree
predictedLikelihood = math.log(0.11279709097648889)
calculatedLikelihood = discrete.mk(tree, charstates, Q)
self.assertTrue(np.isclose(predictedLikelihood, calculatedLikelihood))
def test_mk_fitzjohn_matchesDiversitree(self):
charstates = [
0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0,
0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 1
]
tree = ivy.tree.read("support/randtree100tipsscale2.newick")
Q = np.array([[-2.09613850e-01, 1.204029e-01, 8.921095e-02],
[5.654382e-01, -5.65438217e-01, 1.713339e-08],
[2.415020e-06, 5.958744e-07, -3.01089440e-06]])
expectedLikelihood = -32.79025
calculatedLogLikelihood = discrete.mk(tree,
charstates,
Q,
pi="Fitzjohn")
self.assertTrue(np.isclose(expectedLikelihood,
calculatedLogLikelihood))
def test_mkFlatroot_3tiptreeSymmetricQ2x2_returnslikelihood(self):
tree = self.threetiptree
chars = self.charstates_011
Q = self.Q2x2_sym
# Manually calculated likelihood for expected output
L0A = 1;L1A = 0;L0B = 0;L1B = 1;L0D = 0;L1D = 1
P00A = 0.90936538
P01A = 0.09063462
P11A = 0.90936538
P10A = 0.09063462
P00B = 0.90936538
P01B = 0.09063462
P11B = 0.90936538
P10B = 0.09063462
P00C = 0.90936538
P01C = 0.09063462
P11C = 0.90936538
P10C = 0.09063462
P00D = 0.83516002
P01D = 0.16483998
P11D = 0.83516002
P10D = 0.16483998
L0C = (P00A * L0A + P01A * L1A) * (P00B * L0B + P01B * L1B)
L1C = (P10A * L0A + P11A * L1A) * (P10B * L0B + P11B * L1B)
L0r = (P00C * L0C + P01C * L1C) * (P00D * L0D + P01D * L1D)
L1r = (P10C * L0C + P11C * L1C) * (P10D * L0D + P11D * L1D)
predictedLikelihood = math.log(L0r * 0.5 + L1r * 0.5)
calculatedLikelihood = discrete.mk(tree, chars, Q)
self.assertTrue(np.isclose(predictedLikelihood, calculatedLikelihood))
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 2, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 1]
Q = np.array([[-2.09613850e-01, 1.204029e-01, 8.921095e-02],
[5.654382e-01, -5.65438217e-01, 1.713339e-08],
[2.415020e-06, 5.958744e-07, -3.01089440e-06]])
mk(tree, chars, Q, pi="Equal", returnPi=True)
q, lik = discrete.fitMk(tree, chars, Q="ARD", pi="Equal")
Q = np.array([[-2.09613850e-01, 1.204029e-01, 8.921095e-02],
[5.654382e-01, -5.65438217e-01, 1.713339e-08],
[2.415020e-06, 5.958744e-07, -3.01089440e-06]])
calculatedLikelihood = discrete.mk(tree, chars, Q,
pi = "Fitzjohn")
calculatedLogLikelihood = math.log(calculatedLikelihood)
calculatedLogLikelihood
tree = ivy.tree.read("../support/randtree100tips.newick")
chars = [0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0,
0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
q,l = discrete.fitMk(tree, chars, Q="Sym", pi="Equilibrium")
Q2avg = np.percentile(Q2, 50)
multiregime_Q = np.array([[[-Q1avg,Q1avg],
[Q1avg,-Q1avg]],
[[-Q2avg,Q2avg],
[Q2avg,-Q2avg]]])
multiregime_locs = discrete.locs_from_switchpoint(mr_tree, mr_tree[int(scipy.stats.mode(out["switch"])[0])])
multiregime_likelihood = discrete.mk_mr(mr_tree, mr_chars,
multiregime_Q, multiregime_locs,
pi="Equal")
singleregime_Q = np.array([[-Qavg,Qavg],
[Qavg,-Qavg]])
singleregime_likelihood = discrete.mk(mr_tree, mr_chars, singleregime_Q,
pi="Equal")
multiregime_likelihood
singleregime_likelihood
# Likelihood ratio test: we can do this because the 1-regime model
# is a special case of the 2-regime model
lr = -2*singleregime_likelihood + 2*multiregime_likelihood
# pchisq(19.4675329194382, 1, lower.tail=FALSE)
# 1.023242e-05
# Strong support for two regimes
Q2avg = np.percentile(Q2, 50)
multiregime_Q = np.array([[[-Q1avg,Q1avg],
[Q1avg,-Q1avg]],
[[-Q2avg,Q2avg],
[Q2avg,-Q2avg]]])
multiregime_locs = discrete.locs_from_switchpoint(mr_tree, mr_tree[int(scipy.stats.mode(out["switch"])[0])])
multiregime_likelihood = discrete.mk_multi_regime(mr_tree, mr_chars,
multiregime_Q, multiregime_locs,
pi="Equal")
singleregime_Q = np.array([[-Qavg,Qavg],
[Qavg,-Qavg]])
singleregime_likelihood = discrete.mk(mr_tree, mr_chars, singleregime_Q,
pi="Equal")
multiregime_likelihood
singleregime_likelihood
# Likelihood ratio test: we can do this because the 1-regime model
# is a special case of the 2-regime model
lr = -2*singleregime_likelihood + 2*multiregime_likelihood
# pchisq(19.4675329194382, 1, lower.tail=FALSE)
# 1.023242e-05
# Strong support for two regimes