def setUp(self):
'''Load test data and expected results.'''
unittest.TestCase.setUp(self)
# The way to load a pedigree in conjunction with a genotype set is to recode
# its sample IDs to consecutive for easier access by phasers.
self.problem = io.read_plink(pedigree=itu.HUTT_PED, prefix=itu.GENOTYPE_SAMPLE, haplotype=None)
self.phaser = new_phaser_chain([trivial_phaser()])
def test_phase_family(self):
'''Check phasing trivial cases in all genotyped trios.'''
problem = io.read_plink(pedigree=itu.HUTT_PED, prefix=itu.GENOTYPE_SAMPLE, haplotype=None)
itu.assert_size_equals(problem.genotype, 8, 1415)
assert_equal(len(problem.trios()), 869, 'Unexpected # of genotyped trios')
self.phaser.run(problem, PhaseParam(debug=False))
itu.assert_problem_stats(problem, 22640, 20225, 144)
def problem_family(family, haplotype=False):
'''Load a single nuclear family data set from PLINK data.
If haplotype=True, setting initial haplotypes.'''
return io.read_plink(pedigree=family + '.tfam',
prefix=family,
haplotype=family + '.hap.tped' if haplotype else None,
info=None, idcoef=None, frames=family + '.frm')
def test_estimate_genotype_frequencies(self):
'''Test estimating genotype frequencies for each SNP using the prepare module.'''
problem = io.read_plink(pedigree=itu.HUTT_PED,
prefix=itu.GENOTYPE_SAMPLE,
haplotype=None)
itu.assert_size_equals(problem.genotype, 8, 1415)
self.phaser.run(problem)
# Expected result
frequency = np.array(
[[3.44876319e-01, 4.74911660e-01, 1.75265014e-01, 4.94699646e-03],
[1.06007066e-02, 1.86572433e-01, 8.02826881e-01, 0.00000000e+00],
[1.83745585e-02, 2.93286204e-01, 6.63604259e-01, 2.47349832e-02],
[2.17667848e-01, 5.17314494e-01, 2.64310956e-01, 7.06713763e-04],
[2.16961130e-01, 5.16607761e-01, 2.56537110e-01, 9.89399292e-03],
[6.10600710e-01, 3.33568901e-01, 5.58303893e-02, 0.00000000e+00],
[7.24381626e-01, 1.59010604e-01, 7.06713763e-04, 1.15901060e-01],
[4.19081271e-01, 4.62190807e-01, 1.11660779e-01, 7.06713786e-03]])
assert_almost_equal(problem.info.genotype_frequency,
frequency,
decimal=5,
err_msg='Wrong SNP genotype frequency estimation')
# assert_almost_equal(problem.info.allele_frequency(1), frequency[:, 0] + 0.5 * frequency[:, 1],
# decimal=5, err_msg='Wrong SNP genotype allele frequency estimation')
# assert_almost_equal(problem.info.allele_frequency(2), frequency[:, 2] + 0.5 * frequency[:, 1],
# decimal=5, err_msg='Wrong SNP genotype allele frequency estimation')
assert_almost_equal(
problem.info.allele_frequency(2),
1.0 - problem.info.allele_frequency(1),
decimal=5,
err_msg='Wrong SNP genotype allele frequency estimation')
def setUp(self):
'''Load test data and expected results.'''
unittest.TestCase.setUp(self)
# The way to load a pedigree in conjunction with a genotype set is to recode
# its sample IDs to consecutive for easier access by phasers.
self.problem = io.read_plink(prefix=itu.GENOTYPE_DUO, haplotype=None, pedigree=itu.GENOTYPE_DUO + '.tfam')
self.phaser = trivial_phaser()
# Expected results
self.solution = Problem(self.problem.pedigree, io_genotype.read('plink', 'genotype', prefix=itu.GENOTYPE_DUO_SOLUTION))
def setUp(self):
'''Load test data and expected results.'''
unittest.TestCase.setUp(self)
# The way to load a pedigree in conjunction with a genotype set is to recode
# its sample IDs to consecutive for easier access by phasers.
self.problem = io.read_plink(pedigree=itu.HUTT_PED,
prefix=itu.GENOTYPE_SAMPLE,
haplotype=None)
self.phaser = new_phaser_chain([trivial_phaser()])
def __npz_to_plink(self, p, file_name):
'''Convert p from npz to plink format using the file set specified by file_name.'''
npz = file_name+'.npz'
# Save test problem in plink format
io.write_npz(p, npz)
# Convert plink -> npz
io.npz_to_plink(npz, file_name)
# Load npz and check that the problem object didn't change
p2 = io.read_plink(prefix=file_name)
return p2
def problem_family(family, haplotype=False):
'''Load a single nuclear family data set from PLINK data.
If haplotype=True, setting initial haplotypes.'''
return io.read_plink(pedigree=family + '.tfam',
prefix=family,
haplotype=family +
'.hap.tped' if haplotype else None,
info=None,
idcoef=None,
frames=family + '.frm')
def setUp(self):
'''Load test data and expected results.'''
unittest.TestCase.setUp(self)
# The way to load a pedigree in conjunction with a genotype set is to recode
# its sample IDs to consecutive for easier access by phasers.
self.problem = io.read_plink(prefix=itu.GENOTYPE_TRIO, haplotype=None, pedigree=itu.GENOTYPE_TRIO + '.tfam')
self.phaser = trivial_phaser()
# Expected results
self.solution = Problem(self.problem.pedigree, io_genotype.read('plink', 'genotype', prefix=itu.GENOTYPE_TRIO_SOLUTION))
def __save_and_load_problem_plink(self, problem):
'''Save and load a problem from PLINK file set.'''
try:
# Get a temporary file name
f = tempfile.NamedTemporaryFile(delete=False)
file_name = f.name
f.close()
io.write_plink(problem, file_name)
return io.read_plink(prefix=file_name)
finally:
# Delete test files
for ext in ['', '.pdg.tfam', '.tfam', '.tped', '.hap.tped', '.info']:
os.remove(file_name + ext)
def test_estimate_genotype_frequencies(self):
'''Test estimating genotype frequencies for each SNP using the prepare module.'''
problem = io.read_plink(pedigree=itu.HUTT_PED, prefix=itu.GENOTYPE_SAMPLE, haplotype=None)
itu.assert_size_equals(problem.genotype, 8, 1415)
self.phaser.run(problem)
# Expected result
frequency = np.array([[ 3.44876319e-01, 4.74911660e-01, 1.75265014e-01, 4.94699646e-03],
[ 1.06007066e-02, 1.86572433e-01, 8.02826881e-01, 0.00000000e+00],
[ 1.83745585e-02, 2.93286204e-01, 6.63604259e-01, 2.47349832e-02],
[ 2.17667848e-01, 5.17314494e-01, 2.64310956e-01, 7.06713763e-04],
[ 2.16961130e-01, 5.16607761e-01, 2.56537110e-01, 9.89399292e-03],
[ 6.10600710e-01, 3.33568901e-01, 5.58303893e-02, 0.00000000e+00],
[ 7.24381626e-01, 1.59010604e-01, 7.06713763e-04, 1.15901060e-01],
[ 4.19081271e-01, 4.62190807e-01, 1.11660779e-01, 7.06713786e-03]])
assert_almost_equal(problem.info.genotype_frequency, frequency,
decimal=5, err_msg='Wrong SNP genotype frequency estimation')
# assert_almost_equal(problem.info.allele_frequency(1), frequency[:, 0] + 0.5 * frequency[:, 1],
# decimal=5, err_msg='Wrong SNP genotype allele frequency estimation')
# assert_almost_equal(problem.info.allele_frequency(2), frequency[:, 2] + 0.5 * frequency[:, 1],
# decimal=5, err_msg='Wrong SNP genotype allele frequency estimation')
assert_almost_equal(problem.info.allele_frequency(2), 1.0 - problem.info.allele_frequency(1),
decimal=5, err_msg='Wrong SNP genotype allele frequency estimation')
print usage
sys.exit(util.EXIT_BAD_INPUT_ARGS)
try:
# Prepare file names, create directories
(base_name, pedigree_file, out_base_name) = args
mkdir_if_not_exists(os.path.dirname(out_base_name))
if options.out_gxn:
mkdir_if_not_exists(os.path.dirname(options.out_gxn))
else:
options.out_gxn = out_base_name
npz_file = base_name + '.npz'
# Convert plink tped -> npz
problem = io.read_plink(prefix=base_name, pedigree=pedigree_file, haplotype=None,
verbose=options.debug)
# Phase, impute, fill missing
phaser = phase.build_phasing_pipeline(options)
request = phase.run_phasing_chain(phaser, problem)
stats = request.stats
print ''
stats.pprint()
print ''
# Convert phased npz -> plink tped. Save only genotypes (haplotypes may need to be saved in the stats
# object as a hash table for 'coloring the pedigree' at a later stage.
genotype_file = out_base_name + '.tped'
io.write_plink(problem, out_base_name, verbose=True,
save_node_type=False, save_genotype=True, save_haplotype=False, save_error=False)
