def test_mkMultiRegime_twoQ_matchesByHand(self):
tree = self.threetiptree
chars = self.charstates_011
Q1 = np.array([[-0.1, 0.1], [0.1, -0.1]], dtype=np.double)
Q2 = np.array([[-0.2, 0.2], [0.2, -0.2]], dtype=np.double)
Qs = np.array([Q1, Q2])
inds = np.array([[1, 2, 3], [
4,
]])
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 = 0.72466448
P01D = 0.27533552
P11D = 0.72466448
P10D = 0.27533552
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_multi_regime(tree,
chars,
Qs,
locs=inds,
pi="Equal")
self.assertTrue(np.isclose(predictedLikelihood, calculatedLikelihood))
def test_mkMultiRegime_twoQFourTip_matchesByHand(self):
tree = ivy.tree.read("(((A:1,B:1)C:1,D:2)E:1,F:3)root;")
chars = [0,1,1,0]
Q1 = np.array([[-0.1,0.1],[0.1,-0.1]], dtype=np.double)
Q2 = np.array([[-0.2,0.2],[0.2,-0.2]], dtype=np.double)
Qs = np.array([Q1,Q2])
inds = np.array([[1,2,3],[4,5,6]])
L0A = 1;L1A = 0;L0B = 0;L1B = 1;L0D = 0;L1D = 1;L0F = 1;L1F = 0
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 = 0.72466448
P01D = 0.27533552
P11D = 0.72466448
P10D = 0.27533552
P00E = 0.83516002
P01E = 0.16483998
P11E = 0.83516002
P10E = 0.16483998
P00F = 0.65059711
P01F = 0.34940289
P11F = 0.65059711
P10F = 0.34940289
L0C = (P00A * L0A + P01A * L1A) * (P00B * L0B + P01B * L1B)
L1C = (P10A * L0A + P11A * L1A) * (P10B * L0B + P11B * L1B)
L0E = (P00C * L0C + P01C * L1C) * (P00D * L0D + P01D * L1D)
L1E = (P10C * L0C + P11C * L1C) * (P10D * L0D + P11D * L1D)
L0r = (P00E * L0E + P01E * L1E) * (P00F * L0F + P01F * L1F)
L1r = (P10E * L0E + P11E * L1E) * (P10F * L0F + P11F * L1F)
predictedLikelihood = math.log(L0r * 0.5 + L1r * 0.5)
calculatedLikelihood = discrete.mk_multi_regime(tree, chars, Qs, locs = inds,
pi="Equal")
self.assertTrue(np.isclose(predictedLikelihood, calculatedLikelihood))
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_twoQ_matchesByHand(self):
tree = self.threetiptree
chars = self.charstates_011
Q1 = np.array([[-0.1,0.1],[0.1,-0.1]], dtype=np.double)
Q2 = np.array([[-0.2,0.2],[0.2,-0.2]], dtype=np.double)
Qs = np.array([Q1,Q2])
inds = np.array([[1,2,3],[4,]])
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 = 0.72466448
P01D = 0.27533552
P11D = 0.72466448
P10D = 0.27533552
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_multi_regime(tree, chars, Qs, locs = inds,
pi="Equal")
self.assertTrue(np.isclose(predictedLikelihood, calculatedLikelihood))
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)
# fig = treefig(mr_tree)
# cols = map(lambda x: "black" if x==0 else "red", mr_chars)
#
# fig.add_layer(ivy.vis.layers.add_circles, mr_tree.leaves(),
# colors = cols, size=4)
trueSlowQ = np.array([[-.1,.1], [.1,-.1]])
trueFastQ = np.array([[-.8,.8], [.8,-.8]])
trueQs = np.array([trueSlowQ, trueFastQ])
fastNodes = np.array([ n.ni for n in mr_tree.mrca(mr_tree.grep("f")).preiter()])
slowNodes = np.array([n.ni for n in mr_tree.descendants()if not n.ni in fastNodes])
trueLocs = [fastNodes, slowNodes]
l = discrete.mk_multi_regime(mr_tree, mr_chars, trueQs, trueLocs, pi="Equal")
m = bayesian_models.create_multi_mk_model(mr_tree, mr_chars, Qtype="ER", pi="Equal", nregime=2)
sr_chars = [1, 1, 1, 1, 1, 1, 1, 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, 0, 0, 0, 0, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 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, 1, 1, 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, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
def test_mkMultiRegime_twoQFourTip_matchesByHand(self):
tree = ivy.tree.read("(((A:1,B:1)C:1,D:2)E:1,F:3)root;")
chars = [0, 1, 1, 0]
Q1 = np.array([[-0.1, 0.1], [0.1, -0.1]], dtype=np.double)
Q2 = np.array([[-0.2, 0.2], [0.2, -0.2]], dtype=np.double)
Qs = np.array([Q1, Q2])
inds = np.array([[1, 2, 3], [4, 5, 6]])
L0A = 1
L1A = 0
L0B = 0
L1B = 1
L0D = 0
L1D = 1
L0F = 1
L1F = 0
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 = 0.72466448
P01D = 0.27533552
P11D = 0.72466448
P10D = 0.27533552
P00E = 0.83516002
P01E = 0.16483998
P11E = 0.83516002
P10E = 0.16483998
P00F = 0.65059711
P01F = 0.34940289
P11F = 0.65059711
P10F = 0.34940289
L0C = (P00A * L0A + P01A * L1A) * (P00B * L0B + P01B * L1B)
L1C = (P10A * L0A + P11A * L1A) * (P10B * L0B + P11B * L1B)
L0E = (P00C * L0C + P01C * L1C) * (P00D * L0D + P01D * L1D)
L1E = (P10C * L0C + P11C * L1C) * (P10D * L0D + P11D * L1D)
L0r = (P00E * L0E + P01E * L1E) * (P00F * L0F + P01F * L1F)
L1r = (P10E * L0E + P11E * L1E) * (P10F * L0F + P11F * L1F)
predictedLikelihood = math.log(L0r * 0.5 + L1r * 0.5)
calculatedLikelihood = discrete.mk_multi_regime(tree,
chars,
Qs,
locs=inds,
pi="Equal")
self.assertTrue(np.isclose(predictedLikelihood, calculatedLikelihood))
# fig = treefig(mr_tree)
# cols = map(lambda x: "black" if x==0 else "red", mr_chars)
#
# fig.add_layer(ivy.vis.layers.add_circles, mr_tree.leaves(),
# colors = cols, size=4)
trueSlowQ = np.array([[-.1,.1], [.1,-.1]])
trueFastQ = np.array([[-.8,.8], [.8,-.8]])
trueQs = np.array([trueSlowQ, trueFastQ])
fastNodes = np.array([ n.ni for n in mr_tree.mrca(mr_tree.grep("f")).preiter()])
slowNodes = np.array([n.ni for n in mr_tree.descendants()if not n.ni in fastNodes])
trueLocs = [fastNodes, slowNodes]
l = discrete.mk_multi_regime(mr_tree, mr_chars, trueQs, trueLocs, pi="Equal")
m = bayesian_models.create_multi_mk_model(mr_tree, mr_chars, Qtype="ER", pi="Equal", nregime=2)
mc = pymc.MCMC(m)
mc.sample(5000, burn=200)
out = {"Qparams":mc.trace("Qparams_scaled")[:],
"switch":mc.trace("switch")[:]}
switchids = [ int(i) for i in out["switch"] ]
