def sbool(ped, inds, loc):
npairs = float(len(inds) * (len(inds) - 1) / 2)
r = [
ibs(j.get_genotype(loc, checkhasgeno=False), k.get_genotype(loc, checkhasgeno=False)) > 0
for j, k in itertools.combinations(inds, 2)
]
return sum(r) / npairs
def _writeibd(self, replicatenumber):
# Warning: Don't call this function! If the individuals in the pedigree dont have
# LABEL genotypes, you're just going to get IBS configurations at each locus, not
# actual IBD calculations.
#
# If you have data you want to identify IBD segments in, check
# pydigree.sgs
with smartopen(
'{0}-{1}.ibd.gz'.format(self.label, replicatenumber + 1),
'w') as of:
for ped in self.template.pedigrees:
for ind1, ind2 in combinations_with_replacement(
ped.individuals, 2):
identical = []
for chrom_idx in range(ind1.chromosomes.nchrom()):
if ind1 == ind2:
genos = zip(*ind1.genotypes[chrom_idx])
ibd = [2 * (x == y) for x, y in genos]
else:
genos1 = zip(*ind1.genotypes[chrom_idx])
genos2 = zip(*ind2.genotypes[chrom_idx])
ibd = [
ibs(g1, g2) for g1, g2 in zip(genos1, genos2)
]
identical.extend(ibd)
outline = [ped.label, ind1.label, ind2.label] + identical
outline = ' '.join([str(x) for x in outline])
of.write('{}\n'.format(outline))
def test_ibs():
assert ibs( (2,2), (2,2) ) == 2
assert ibs( (1,2), (1,2) ) == 2
assert ibs( (2,1), (2,2) ) == 1
assert ibs( (2,2), (2,1) ) == 1
assert ibs( (1,1), (2,2) ) == 0
assert ibs( (1,1), (0,0), missingval=64) == 64
assert ibs( (1,1), (0,0) ) == None
assert ibs( (0,0), (1,1) ) == None
def spairs(inds, loc):
"Returns total number of IBD alleles"
r = [ibs(j.get_genotype(loc, checkhasgeno=False),
k.get_genotype(loc, checkhasgeno=False))
for j, k in itertools.combinations(inds, 2)]
return sum(r)
def spairs(inds, loc):
"Returns total number of IBD alleles"
r = [
ibs(j.get_genotype(loc, checkhasgeno=False),
k.get_genotype(loc, checkhasgeno=False))
for j, k in itertools.combinations(inds, 2)
]
return sum(r)
def sbool(inds, loc):
"Returns proportion of pairs IBD != 0"
npairs = float(len(inds) * (len(inds) - 1) / 2)
r = [
ibs(j.get_genotype(loc, checkhasgeno=False),
k.get_genotype(loc, checkhasgeno=False)) > 0
for j, k in itertools.combinations(inds, 2)
]
return sum(r) / npairs
def _writeibd(self, replicatenumber):
# Warning: Don't call this function! If the individuals in the pedigree dont have
# LABEL genotypes, you're just going to get IBS configurations at each locus, not
# actual IBD calculations.
#
# If you have data you want to identify IBD segments in, check
# pydigree.sgs
with smartopen('{0}-{1}.ibd.gz'.format(self.label, replicatenumber + 1), 'w') as of:
for ped in self.template.pedigrees:
for ind1, ind2 in combinations_with_replacement(ped.individuals, 2):
identical = []
for chrom_idx in range(ind1.chromosomes.nchrom()):
if ind1 == ind2:
genos = zip(*ind1.genotypes[chrom_idx])
ibd = [2 * (x == y) for x, y in genos]
else:
genos1 = zip(*ind1.genotypes[chrom_idx])
genos2 = zip(*ind2.genotypes[chrom_idx])
ibd = [ibs(g1, g2)
for g1, g2 in zip(genos1, genos2)]
identical.extend(ibd)
outline = [ped.label, ind1.label, ind2.label] + identical
outline = ' '.join([str(x) for x in outline])
of.write('{}\n'.format(outline))
def spairs(ped, inds, loc):
r = [
ibs(j.get_genotype(loc, checkhasgeno=False), k.get_genotype(loc, checkhasgeno=False))
for j, k in itertools.combinations(inds, 2)
]
return sum(r)
s = pyd.sgs.sgs_population(ped, seed_size=ms)
with smartopen('{}-{}.ibd.gz'.format(prefix, replicate)) as f:
trueibd = {}
for line in f:
fam, id1, id2, ibd_states = line.strip().split(None, 3)
trueibd[frozenset({id1,id2})] = np.array([int(x) for x in ibd_states.split()])
a = intervals_to_array(s[frozenset({ped['7'],ped['8']})][0], ped.chromosomes[0].nmark())
b = trueibd[frozenset({'7','8'})]
genos1 = zip(*ped['7'].genotypes[0])
genos2 = zip(*ped['8'].genotypes[0])
identical = [ibs(x,y) for x,y in zip(genos1, genos2)]
from pydigree.common import table, runs
for start, stop in runs(list(a), lambda x: x>0, ms):
print('Predicted segment: {}-{}'.format(start, stop))
print()
for start, stop in runs(list(b), lambda x: x>0, 2):
print('True IBD Segment: {}-{}'.format(start, stop))
correct_calls = a == b
print('Accuracy: {}'.format(correct_calls.sum() / float(correct_calls.shape[0])))