def test_Fst__windowed(sample_size, n_cohorts, chunks):
ts = simulate_ts(sample_size, length=200)
ds = ts_to_dataset(ts, chunks) # type: ignore[no-untyped-call]
ds, subsets = add_cohorts(ds, ts,
n_cohorts) # type: ignore[no-untyped-call]
ds = window(ds, size=25)
fst_ds = Fst(ds, estimator="Nei")
fst = fst_ds["stat_Fst"].values
# Calculate Fst using tskit windows
# Find the variant positions so we can have windows with a fixed number of variants
positions = ts.tables.sites.position
windows = np.concatenate(([0], positions[::25][1:], [ts.sequence_length]))
n_windows = len(windows) - 1
ts_fst = np.full([n_windows, n_cohorts, n_cohorts], np.nan)
for i, j in itertools.combinations(range(n_cohorts), 2):
ts_fst[:, i, j] = ts.Fst([subsets[i], subsets[j]],
windows=windows,
span_normalise=False)
ts_fst[:, j, i] = ts_fst[:, i, j]
# We can values close to zero, and the default value of atol isn't
# appropriate for this.
atol = 1e-8
np.testing.assert_allclose(fst, ts_fst, atol=atol)
# scikit-allel
fst_ds = Fst(ds, estimator="Hudson")
for i, j in itertools.combinations(range(n_cohorts), 2):
fst = fst_ds["stat_Fst"].sel(cohorts_0=f"co_{i}",
cohorts_1=f"co_{j}").values
ac_i = fst_ds.cohort_allele_count.values[:, i, :]
ac_j = fst_ds.cohort_allele_count.values[:, j, :]
ska_fst = allel.moving_hudson_fst(ac_i, ac_j, size=25)
np.testing.assert_allclose(
fst[:-1], ska_fst,
atol=atol) # scikit-allel has final window missing
def test_Fst__Nei(sample_size, n_cohorts):
ts = simulate_ts(sample_size)
ds = ts_to_dataset(ts)
ds, subsets = add_cohorts(ds, ts, n_cohorts)
n_variants = ds.dims["variants"]
ds = window_by_variant(ds, size=n_variants) # single window
ds = Fst(ds, estimator="Nei")
fst = ds.stat_Fst.values
ts_fst = np.full([1, n_cohorts, n_cohorts], np.nan)
for i, j in itertools.combinations(range(n_cohorts), 2):
ts_fst[0, i, j] = ts.Fst([subsets[i], subsets[j]])
ts_fst[0, j, i] = ts_fst[0, i, j]
np.testing.assert_allclose(fst, ts_fst)
def test_Fst__Hudson(sample_size):
# scikit-allel can only calculate Fst for pairs of cohorts (populations)
n_cohorts = 2
ts = simulate_ts(sample_size)
ds = ts_to_dataset(ts)
ds, subsets = add_cohorts(ds, ts, n_cohorts)
n_variants = ds.dims["variants"]
ds = window_by_variant(ds, size=n_variants) # single window
ds = Fst(ds, estimator="Hudson")
fst = ds.stat_Fst.sel(cohorts_0="co_0", cohorts_1="co_1").values
# scikit-allel
ac1 = ds.cohort_allele_count.values[:, 0, :]
ac2 = ds.cohort_allele_count.values[:, 1, :]
num, den = hudson_fst(ac1, ac2)
ska_fst = np.sum(num) / np.sum(den)
np.testing.assert_allclose(fst, ska_fst)
def test_Fst__Nei(sample_size, n_cohorts):
ts = msprime.simulate(sample_size,
length=100,
mutation_rate=0.05,
random_seed=42)
ds = ts_to_dataset(ts) # type: ignore[no-untyped-call]
ds, subsets = add_cohorts(ds, ts,
n_cohorts) # type: ignore[no-untyped-call]
n_variants = ds.dims["variants"]
ds = window(ds, size=n_variants) # single window
ds = Fst(ds, estimator="Nei")
fst = ds.stat_Fst.values
ts_fst = np.full([1, n_cohorts, n_cohorts], np.nan)
for i, j in itertools.combinations(range(n_cohorts), 2):
ts_fst[0, i, j] = ts.Fst([subsets[i], subsets[j]])
ts_fst[0, j, i] = ts_fst[0, i, j]
np.testing.assert_allclose(fst, ts_fst)
def test_Fst__Hudson(sample_size):
# scikit-allel can only calculate Fst for pairs of cohorts (populations)
n_cohorts = 2
ts = msprime.simulate(sample_size,
length=100,
mutation_rate=0.05,
random_seed=42)
ds = ts_to_dataset(ts) # type: ignore[no-untyped-call]
ds, subsets = add_cohorts(ds, ts,
n_cohorts) # type: ignore[no-untyped-call]
n_variants = ds.dims["variants"]
ds = window(ds, size=n_variants) # single window
ds = Fst(ds, estimator="Hudson")
fst = ds.stat_Fst.sel(cohorts_0="co_0", cohorts_1="co_1").values
# scikit-allel
ac1 = ds.cohort_allele_count.values[:, 0, :]
ac2 = ds.cohort_allele_count.values[:, 1, :]
num, den = hudson_fst(ac1, ac2)
ska_fst = np.sum(num) / np.sum(den)
np.testing.assert_allclose(fst, ska_fst)
def test_Fst__unknown_estimator():
ts = simulate_ts(2)
ds = ts_to_dataset(ts) # type: ignore[no-untyped-call]
with pytest.raises(ValueError,
match="Estimator 'Unknown' is not a known estimator"):
Fst(ds, estimator="Unknown")
def test_Fst__unknown_estimator():
ts = msprime.simulate(2, length=100, mutation_rate=0.05, random_seed=42)
ds = ts_to_dataset(ts) # type: ignore[no-untyped-call]
with pytest.raises(ValueError,
match="Estimator 'Unknown' is not a known estimator"):
Fst(ds, estimator="Unknown")
