def verify(
self,
tree,
newick=None,
initial_size=1,
branch_length_units="gen",
generation_time=None,
):
if newick is None:
newick = self.make_newick(tree)
demography = species_trees.parse_species_tree(
newick,
initial_size=initial_size,
branch_length_units=branch_length_units,
generation_time=generation_time,
)
assert demography.num_populations == tree.num_nodes
for pop in demography.populations:
assert pop.initial_size == initial_size
assert pop.growth_rate == 0
assert pop.name is not None
# Population IDs are mapped to leaves first, and then to the internal nodes
# in postorder
pop_id_map = {}
k = 0
for u in tree.leaves():
pop_id_map[u] = k
k += 1
for u in tree.nodes(order="postorder"):
if tree.is_internal(u):
pop_id_map[u] = k
k += 1
for u in tree.nodes():
pop = demography.populations[pop_id_map[u]]
assert pop.growth_rate == 0
if tree.is_leaf(u):
# Assuming we're using the make_newick function above
assert pop.name == f"node_{u}"
# We should have demographic events for every internal node, and
# events should be output in increasing time order.
j = 0
for node in tree.nodes(order="timeasc"):
children = tree.children(node)
dest = pop_id_map[node]
for child in children:
event = demography.events[j]
j += 1
assert isinstance(event, msprime.MassMigration)
assert event.time == pytest.approx(tree.time(node))
source = pop_id_map[child]
assert event.source == source
assert event.dest == dest
assert j == len(demography.events)
def test_bad_parameter(self):
good_tree = "(((human:5.6,chimpanzee:5.6):3.0,gorilla:8.6):9.4,orangutan:18.0)"
good_time_units = "myr"
good_ne = 10000
good_generation_time = 5
for bad_time_units in [-3, "asdf", ["myr"]]:
with pytest.raises(ValueError):
species_trees.parse_species_tree(
good_tree,
time_units=bad_time_units,
initial_size=good_ne,
generation_time=good_generation_time,
)
with pytest.raises(TypeError):
species_trees.parse_species_tree(good_tree, None)
for bad_ne in [-3, "x"]:
with pytest.raises(ValueError):
species_trees.parse_species_tree(
good_tree,
time_units=good_time_units,
initial_size=bad_ne,
generation_time=good_generation_time,
)
for bad_generation_time in [None, -3, "x"]:
with pytest.raises(ValueError):
species_trees.parse_species_tree(
good_tree,
time_units=good_time_units,
initial_size=good_ne,
generation_time=bad_generation_time,
)
for bad_time_units in ["gen"]:
with pytest.raises(ValueError):
species_trees.parse_species_tree(
good_tree,
time_units=bad_time_units,
initial_size=good_ne,
generation_time=good_generation_time,
)
def test_bad_tree(self):
bad_trees = [
"",
";",
"abcd",
";;;",
"___",
"∞",
"(",
")",
"()",
"( )",
"(()())",
"((3:0.39,5:0.39]:1.39,(4:0.47,(1:0.18,2:0.18):0.29):1.31);",
"((3:0.39,5:0.39(:1.39,(4:0.47,(1:0.18,2:0.18):0.29):1.31);",
"((3:0.39,5:0.39,:1.39,(4:0.47,(1:0.18,2:0.18):0.29):1.31);",
"(4:0.47,(1:0.18,2:0.18):0.29):1.31);",
]
for bad_tree in bad_trees:
with pytest.raises(ValueError):
species_trees.parse_species_tree(tree=bad_tree, initial_size=1)
def test_bad_params(self):
with pytest.raises(TypeError):
species_trees.parse_species_tree()
with pytest.raises(TypeError):
species_trees.parse_species_tree(tree="()")
with pytest.raises(TypeError):
species_trees.parse_species_tree(initial_size=1)
def test_4_species_run(self):
species_tree = (
"(((human:5.6,chimpanzee:5.6):3.0,gorilla:8.6):9.4,orangutan:18.0)"
)
spec = species_trees.parse_species_tree(
species_tree,
time_units="myr",
initial_size=10000,
generation_time=20,
)
# Take one sample from each population
ts = msprime.sim_ancestry(samples={j: 1
for j in range(4)},
demography=spec,
ploidy=1)
assert ts.num_trees == 1
assert ts.num_samples == 4
assert ts.num_populations == 7
for j, u in enumerate(ts.samples()):
assert ts.node(u).population == j
pops = list(ts.populations())
assert pops[0].metadata["name"] == "human"
assert pops[1].metadata["name"] == "chimpanzee"
assert pops[2].metadata["name"] == "gorilla"
assert pops[3].metadata["name"] == "orangutan"
assert pops[4].metadata["name"] == "pop_4"
assert pops[5].metadata["name"] == "pop_5"
assert pops[6].metadata["name"] == "pop_6"
# Use the population names to get the samples
samples = dict(human=4, gorilla=2)
ts = msprime.sim_ancestry(samples=samples, demography=spec)
assert ts.num_trees == 1
assert ts.num_samples == 12
for j, u in enumerate(ts.samples()):
pop = 0 if j < 8 else 2
assert ts.node(u).population == pop
# Order of keywords is respected
ts = msprime.sim_ancestry(samples={
"gorilla": 2,
"human": 4
},
demography=spec)
assert ts.num_trees == 1
assert ts.num_samples == 12
for j, u in enumerate(ts.samples()):
pop = 2 if j < 4 else 0
assert ts.node(u).population == pop
def test_4_species_parse(self):
good_tree = "(((human:5.6,chimpanzee:5.6):3.0,gorilla:8.6):9.4,orangutan:18.0)"
good_time_units = "myr"
good_ne = 10000
good_generation_time = 20
demography = species_trees.parse_species_tree(
good_tree,
time_units=good_time_units,
initial_size=good_ne,
generation_time=good_generation_time,
)
assert isinstance(demography.populations, list)
assert len(demography.populations) == 7
for pop in demography.populations:
assert isinstance(pop, msprime.demography.Population)
assert isinstance(demography.events, list)
assert len(demography.events) == 3
for mm in demography.events:
assert isinstance(mm, msprime.demography.PopulationSplit)
def test_duplicate_name(self):
with pytest.raises(ValueError, match="Duplicate population name"):
species_trees.parse_species_tree(tree="(popA:100.0,popA:100.0)",
initial_size=1)
def test_unequal_branch_lengths(self):
with pytest.raises(ValueError):
species_trees.parse_species_tree(tree="(popA:100.0,popB:10.0)",
initial_size=1000)
def verify_non_ultrametric(self, tree):
newick = tree.newick()
with pytest.raises(ValueError):
species_trees.parse_species_tree(newick, initial_size=1)
