def test_pool(self):
nreps = 20000
fm = frag.FragmentationModel(seed=1234,
number_of_fragments=5,
number_of_genomes_per_fragment=10,
generations_since_fragmentation=1000,
effective_pop_size_of_fragment=1000,
effective_pop_size_of_ancestor=10000,
mutation_rate=1e-6,
migration_rate=0.0)
pi, pi_a, pi_w = frag.run_mspi_simulations(fm,
number_of_processes=None,
number_of_replicates=nreps)
self.assertEqual(len(pi), nreps)
self.assertEqual(len(pi_a), nreps)
self.assertEqual(len(pi_w), nreps)
mean_pi = sum(pi) / len(pi)
mean_pi_a = sum(pi_a) / len(pi_a)
mean_pi_w = sum(pi_w) / len(pi_w)
print("")
print(mean_pi)
print(fm.expected_divergence)
print(mean_pi_a)
print(fm.expected_divergence_between_fragments)
print(mean_pi_w)
print(fm.expected_divergence_within_fragments)
self.assertAlmostEqual(fm.expected_divergence, mean_pi, places=3)
self.assertAlmostEqual(fm.expected_divergence_between_fragments,
mean_pi_a,
places=3)
self.assertAlmostEqual(fm.expected_divergence_within_fragments,
mean_pi_w,
places=3)
def test_mspi_expected_divergence_within(self):
fm = frag.FragmentationModel(seed=12345678,
number_of_fragments=2,
number_of_genomes_per_fragment=10,
generations_since_fragmentation=10000,
effective_pop_size_of_fragment=10000,
effective_pop_size_of_ancestor=100000,
mutation_rate=1e-6,
migration_rate=0.0)
self.assertEqual(fm.number_of_fragments, 2)
pi, pi_a, pi_w = fm.mspi_simulate(number_of_replicates=100000)
mean_pi = sum(pi) / len(pi)
mean_pi_a = sum(pi_a) / len(pi_a)
mean_pi_w = sum(pi_w) / len(pi_w)
print("")
print(mean_pi)
print(fm.expected_divergence)
print(mean_pi_a)
print(fm.expected_divergence_between_fragments)
print(mean_pi_w)
print(fm.expected_divergence_within_fragments)
self.assertAlmostEqual(fm.expected_divergence, mean_pi, places=2)
self.assertAlmostEqual(fm.expected_divergence_between_fragments,
mean_pi_a,
places=2)
self.assertAlmostEqual(fm.expected_divergence_within_fragments,
mean_pi_w,
places=2)
def test_number_of_fragments(self):
gens_since_frag = 20000.0
pop_size = 10000
mutation_rate = 1e-6
expected_div = frag.get_expected_divergence_between_fragments(
generations_since_fragmentation=gens_since_frag,
effective_pop_size_of_ancestor=pop_size,
mutation_rate=mutation_rate)
fm = frag.FragmentationModel(
seed=12345678,
number_of_fragments=10,
number_of_genomes_per_fragment=1,
generations_since_fragmentation=gens_since_frag,
effective_pop_size_of_fragment=1.0, # this shouldn't matter
effective_pop_size_of_ancestor=pop_size,
mutation_rate=mutation_rate,
migration_rate=0.0)
self.assertEqual(fm.number_of_fragments, 10)
# pi_summary = stats.SampleSummarizer(
# fm.sample_pi(400000))
pi, pi_a, pi_w = fm.mspi_simulate(number_of_replicates=400000)
mean_pi_a = sum(pi_a) / len(pi_a)
mean_pi = sum(pi) / len(pi)
self.assertAlmostEqual(expected_div, mean_pi_a, places=3)
self.assertAlmostEqual(mean_pi, mean_pi_a, places=3)
def test_mspi_expected_divergence_with_one_sample(self):
gens_since_frag = 2000.0
pop_size = 10000
mutation_rate = 1e-6
expected_div, d_b, d_w = frag.get_expected_divergence(
number_of_fragments=2,
number_of_genomes_per_fragment=1,
generations_since_fragmentation=gens_since_frag,
effective_pop_size_of_ancestor=pop_size,
effective_pop_size_of_fragment=1000,
mutation_rate=mutation_rate)
fm = frag.FragmentationModel(
seed=12345,
number_of_fragments=2,
number_of_genomes_per_fragment=1,
generations_since_fragmentation=gens_since_frag,
effective_pop_size_of_fragment=1000,
effective_pop_size_of_ancestor=pop_size,
mutation_rate=mutation_rate,
migration_rate=0.0)
self.assertEqual(fm.number_of_fragments, 2)
pi, pi_a, pi_w = fm.mspi_simulate(number_of_replicates=50000)
mean_pi = sum(pi) / len(pi)
mean_pi_a = sum(pi_a) / len(pi_a)
print("")
print(mean_pi)
print(expected_div)
print(mean_pi_a)
print(d_b)
self.assertAlmostEqual(expected_div, mean_pi, places=3)
self.assertAlmostEqual(expected_div, d_b, places=3)
self.assertTrue(d_w is None)