def hudson_fst(ac1, ac2, fill=np.nan):
"""Calculate the numerator and denominator for Fst estimation using the
method of Hudson (1992) elaborated by Bhatia et al. (2013).
Parameters
----------
ac1 : array_like, int, shape (n_variants, n_alleles)
Allele counts array from the first population.
ac2 : array_like, int, shape (n_variants, n_alleles)
Allele counts array from the second population.
fill : float
Use this value where there are no pairs to compare (e.g.,
all allele calls are missing).
Returns
-------
num : ndarray, float, shape (n_variants,)
Divergence between the two populations minus average
of diversity within each population.
den : ndarray, float, shape (n_variants,)
Divergence between the two populations.
Examples
--------
Calculate numerator and denominator for Fst estimation::
>>> import allel
>>> g = allel.GenotypeArray([[[0, 0], [0, 0], [1, 1], [1, 1]],
... [[0, 1], [0, 1], [0, 1], [0, 1]],
... [[0, 0], [0, 0], [0, 0], [0, 0]],
... [[0, 1], [1, 2], [1, 1], [2, 2]],
... [[0, 0], [1, 1], [0, 1], [-1, -1]]])
>>> subpops = [[0, 1], [2, 3]]
>>> ac1 = g.count_alleles(subpop=subpops[0])
>>> ac2 = g.count_alleles(subpop=subpops[1])
>>> num, den = allel.hudson_fst(ac1, ac2)
>>> num
array([ 1. , -0.16666667, 0. , -0.125 , -0.33333333])
>>> den
array([1. , 0.5 , 0. , 0.625, 0.5 ])
Estimate Fst for each variant individually::
>>> fst = num / den
>>> fst
array([ 1. , -0.33333333, nan, -0.2 , -0.66666667])
Estimate Fst averaging over variants::
>>> fst = np.sum(num) / np.sum(den)
>>> fst
0.1428571428571429
""" # flake8: noqa
# check inputs
ac1 = asarray_ndim(ac1, 2)
ac2 = asarray_ndim(ac2, 2)
check_dim0_aligned(ac1, ac2)
ac1, ac2 = ensure_dim1_aligned(ac1, ac2)
# calculate these once only
an1 = np.sum(ac1, axis=1)
an2 = np.sum(ac2, axis=1)
# calculate average diversity (a.k.a. heterozygosity) within each
# population
within = (mean_pairwise_difference(ac1, an1, fill=fill) +
mean_pairwise_difference(ac2, an2, fill=fill)) / 2
# calculate divergence (a.k.a. heterozygosity) between each population
between = mean_pairwise_difference_between(ac1, ac2, an1, an2, fill=fill)
# define numerator and denominator for Fst calculations
num = between - within
den = between
return num, den
def hudson_fst(ac1, ac2, fill=np.nan):
"""Calculate the numerator and denominator for Fst estimation using the
method of Hudson (1992) elaborated by Bhatia et al. (2013).
Parameters
----------
ac1 : array_like, int, shape (n_variants, n_alleles)
Allele counts array from the first population.
ac2 : array_like, int, shape (n_variants, n_alleles)
Allele counts array from the second population.
fill : float
Use this value where there are no pairs to compare (e.g.,
all allele calls are missing).
Returns
-------
num : ndarray, float, shape (n_variants,)
Divergence between the two populations minus average
of diversity within each population.
den : ndarray, float, shape (n_variants,)
Divergence between the two populations.
Examples
--------
Calculate numerator and denominator for Fst estimation::
>>> import allel
>>> g = allel.GenotypeArray([[[0, 0], [0, 0], [1, 1], [1, 1]],
... [[0, 1], [0, 1], [0, 1], [0, 1]],
... [[0, 0], [0, 0], [0, 0], [0, 0]],
... [[0, 1], [1, 2], [1, 1], [2, 2]],
... [[0, 0], [1, 1], [0, 1], [-1, -1]]])
>>> subpops = [[0, 1], [2, 3]]
>>> ac1 = g.count_alleles(subpop=subpops[0])
>>> ac2 = g.count_alleles(subpop=subpops[1])
>>> num, den = allel.stats.hudson_fst(ac1, ac2)
>>> num
array([ 1. , -0.16666667, 0. , -0.125 , -0.33333333])
>>> den
array([ 1. , 0.5 , 0. , 0.625, 0.5 ])
Estimate Fst for each variant individually::
>>> fst = num / den
>>> fst
array([ 1. , -0.33333333, nan, -0.2 , -0.66666667])
Estimate Fst averaging over variants::
>>> fst = np.sum(num) / np.sum(den)
>>> fst
0.1428571428571429
""" # flake8: noqa
# check inputs
ac1 = asarray_ndim(ac1, 2)
ac2 = asarray_ndim(ac2, 2)
check_dim0_aligned(ac1, ac2)
ac1, ac2 = ensure_dim1_aligned(ac1, ac2)
# calculate these once only
an1 = np.sum(ac1, axis=1)
an2 = np.sum(ac2, axis=1)
# calculate average diversity (a.k.a. heterozygosity) within each
# population
within = (mean_pairwise_difference(ac1, an1, fill=fill) +
mean_pairwise_difference(ac2, an2, fill=fill)) / 2
# calculate divergence (a.k.a. heterozygosity) between each population
between = mean_pairwise_difference_between(ac1, ac2, an1, an2, fill=fill)
# define numerator and denominator for Fst calculations
num = between - within
den = between
return num, den