def test_phenos_agree_on_features():
constraints = {
"d0": {
"must_be_same": True,
"method": "equal",
"params": {}
},
"d1": {
"must_be_same": False,
"method": "range",
"params": {
"low": 1,
"high": 2
},
},
}
pheno_compare = phenotype_compare.PhenotypeCompare(constraints)
p1 = {"d0": True, "d1": 1.5}
p2 = {"d0": True, "d1": 1.5}
p3 = {"d0": True, "d1": 0.5}
p4 = {"d0": False, "d1": 1.5}
features = {"d0", "d1"}
assert pheno_compare.phenos_agree_on_features(p1, p2, features)
assert not pheno_compare.phenos_agree_on_features(p1, p3, features)
assert not pheno_compare.phenos_agree_on_features(p1, p4, features)
def run(options):
with open(options.case_names_file) as f:
case_sample_names = [x.rstrip() for x in f]
genos = genotypes.Genotypes(
file_of_vcf_filenames=options.vcfs_tsv,
distance_matrix_file=options.distance_matrix,
variant_counts_file=options.var_counts_file,
)
phenos = phenotypes.Phenotypes(options.phenos_tsv)
with open(options.pheno_constraints_json) as f:
pheno_constraints = json.load(f)
pheno_compare = phenotype_compare.PhenotypeCompare(pheno_constraints)
triples = strain_triples.StrainTriples(
genos,
phenos,
pheno_compare,
top_n_genos=options.top_n_genos,
max_pheno_diffs=options.max_pheno_diffs,
processes=options.processes,
)
triples.run_analysis(case_sample_names,
options.out,
mask_file=options.mask_bed_file)
def test_satisfy_required_differences_two_the_same():
constraints = {
"d0": {
"must_be_same": True,
"method": "equal",
"params": {}
},
"d1": {
"must_be_same": True,
"method": "equal",
"params": {}
},
"d2": {
"must_be_same": False,
"method": "range",
"params": {
"low": 1,
"high": 2
},
},
}
pheno_compare = phenotype_compare.PhenotypeCompare(constraints)
p1 = {"d0": True, "d1": True, "d2": 1.5}
p2 = {"d0": True, "d1": True, "d2": 1.1}
p3 = {"d0": True, "d1": False, "d2": 1.5}
p4 = {"d0": True, "d1": True, "d2": 1.5}
p5 = {"d0": True, "d1": True, "d2": 2.5}
assert not pheno_compare.satisfy_required_differences(p1, p2)
assert not pheno_compare.satisfy_required_differences(p2, p1)
assert not pheno_compare.satisfy_required_differences(p1, p3)
assert not pheno_compare.satisfy_required_differences(p3, p1)
assert not pheno_compare.satisfy_required_differences(p1, p4)
assert not pheno_compare.satisfy_required_differences(p4, p1)
assert pheno_compare.satisfy_required_differences(p1, p5)
assert pheno_compare.satisfy_required_differences(p5, p1)
def test_satisfy_required_differences_one_the_same():
constraints = {
"d0": {
"must_be_same": True,
"method": "equal",
"params": {}
},
"d1": {
"must_be_same": False,
"method": "equal",
"params": {}
},
}
pheno_compare = phenotype_compare.PhenotypeCompare(constraints)
p1 = {"d1": True}
p2 = {"d1": True}
p3 = {"d1": False}
p4 = {"d1": None}
p5 = {"d1": None}
assert pheno_compare.satisfy_required_differences(p1, p3)
assert pheno_compare.satisfy_required_differences(p3, p1)
assert not pheno_compare.satisfy_required_differences(p1, p1)
assert not pheno_compare.satisfy_required_differences(p1, p2)
assert not pheno_compare.satisfy_required_differences(p2, p1)
assert not pheno_compare.satisfy_required_differences(p1, p4)
assert not pheno_compare.satisfy_required_differences(p4, p1)
assert not pheno_compare.satisfy_required_differences(p4, p5)
assert not pheno_compare.satisfy_required_differences(p5, p4)
def run(options):
wanted_phenos = utils.command_line_wanted_phenos_to_dict(
options.wanted_pheno)
phenos = phenotypes.Phenotypes(options.phenos_tsv)
with open(options.pheno_constraints_json) as f:
pheno_constraints = json.load(f)
pheno_compare = phenotype_compare.PhenotypeCompare(pheno_constraints)
samples = phenos.find_matching_cases(wanted_phenos, pheno_compare)
if len(samples) == 0:
raise RuntimeError("No matching samples found")
with open(options.outfile, "w") as f:
print(*samples, sep="\n", file=f)
def test_ranked_neighbours_for_one_sample(genos, phenos, constraints):
pheno_compare = phenotype_compare.PhenotypeCompare(constraints)
got = sample_neighbours_finding.ranked_neighbours_for_one_sample(
genos, phenos, pheno_compare, "s1")
expect = [
sample_neighbours_finding.RankData(
sample="s5",
rank_sum=2,
geno_rank=0,
pheno_rank=2,
geno_dist=1,
pheno_dist=3,
),
sample_neighbours_finding.RankData(
sample="s4",
rank_sum=2,
geno_rank=1,
pheno_rank=1,
geno_dist=2,
pheno_dist=2,
),
sample_neighbours_finding.RankData(
sample="s3",
rank_sum=2,
geno_rank=2,
pheno_rank=0,
geno_dist=3,
pheno_dist=1,
),
]
assert got == expect
got = sample_neighbours_finding.ranked_neighbours_for_one_sample(
genos, phenos, pheno_compare, "s1", max_pheno_dist=2)
del expect[0]
assert got == expect
got = sample_neighbours_finding.ranked_neighbours_for_one_sample(
genos, phenos, pheno_compare, "s1", top_n_genos=2)
expect = [
sample_neighbours_finding.RankData(
sample="s5",
rank_sum=0,
geno_rank=0,
pheno_rank=0,
geno_dist=1,
pheno_dist=3,
)
]
assert got == expect
def test_differences():
constraints = {
"d0": {
"must_be_same": True,
"method": "equal",
"params": {}
},
"d1": {
"must_be_same": True,
"method": "equal",
"params": {}
},
"d2": {
"must_be_same": True,
"method": "equal",
"params": {}
},
"d3": {
"must_be_same": False,
"method": "range",
"params": {
"low": 1,
"high": 2
},
},
}
pheno_compare = phenotype_compare.PhenotypeCompare(constraints)
p1 = {"d0": True, "d1": True, "d2": True, "d3": 2.5}
p2 = {"d0": True, "d1": True, "d2": True, "d3": 2.5}
p3 = {"d0": True, "d1": True, "d2": False, "d3": 2.5}
p4 = {"d0": True, "d1": False, "d2": False, "d3": 2.5}
p5 = {"d0": False, "d1": False, "d2": False, "d3": 2.5}
assert pheno_compare.differences(p1, p1) == 0
assert pheno_compare.differences(p1, p2) == 0
assert pheno_compare.differences(p2, p1) == 0
assert pheno_compare.differences(p1, p3) == 1
assert pheno_compare.differences(p3, p1) == 1
assert pheno_compare.differences(p1, p4) == 2
assert pheno_compare.differences(p4, p1) == 2
assert pheno_compare.differences(p1, p5) == 3
assert pheno_compare.differences(p5, p1) == 3
def test_run_analysis(genos, phenos, constraints, caplog):
caplog.set_level(logging.INFO)
pheno_compare = phenotype_compare.PhenotypeCompare(constraints)
triples = strain_triples.StrainTriples(genos,
phenos,
pheno_compare,
top_n_genos=10)
case_sample_names = ["s1", "s2"]
tmp_dir = "tmp.strain_triples.run_analysis"
subprocess.check_output(f"rm -rf {tmp_dir}", shell=True)
os.mkdir(tmp_dir)
outprefix = os.path.join(tmp_dir, "out")
mask_file = os.path.join(data_dir, "run_analysis.mask.bed")
got = triples.run_analysis(case_sample_names,
outprefix,
mask_file=mask_file)
# The contents of these files are checked elsewhere. This test is just to
# check then run of whole pipeline doesn't crash
for filename in got.values():
assert os.path.exists(filename)
subprocess.check_output(f"rm -r {tmp_dir}", shell=True)
def test_find_strain_triples(genos, phenos, constraints, caplog):
caplog.set_level(logging.INFO)
pheno_compare = phenotype_compare.PhenotypeCompare(constraints)
triples = strain_triples.StrainTriples(genos,
phenos,
pheno_compare,
top_n_genos=10)
case_sample_names = ["s1", "s2"]
got_triples = triples.find_strain_triples(case_sample_names)
rank_data_s1_1 = sample_neighbours_finding.RankData(sample="s3",
rank_sum=0,
geno_rank=0,
pheno_rank=0,
geno_dist=1,
pheno_dist=0)
rank_data_s1_2 = sample_neighbours_finding.RankData(sample="s4",
rank_sum=0,
geno_rank=0,
pheno_rank=0,
geno_dist=1,
pheno_dist=0)
rank_data_s2_1 = sample_neighbours_finding.RankData(sample="s6",
rank_sum=0,
geno_rank=0,
pheno_rank=0,
geno_dist=1,
pheno_dist=0)
rank_data_s2_2 = sample_neighbours_finding.RankData(sample="s7",
rank_sum=1,
geno_rank=1,
pheno_rank=0,
geno_dist=2,
pheno_dist=0)
expect = [
strain_triple.StrainTriple("s1", rank_data_s1_1, rank_data_s1_2),
strain_triple.StrainTriple("s2", rank_data_s2_1, rank_data_s2_2),
]
assert got_triples == expect
def test_geno_and_pheno_distances_for_one_sample(genos, phenos, constraints):
pheno_compare = phenotype_compare.PhenotypeCompare(constraints)
(
got_geno_distances,
got_pheno_distances,
) = sample_neighbours_finding._geno_and_pheno_distances_for_one_sample(
genos, phenos, pheno_compare, "s1")
expect_geno_distances = {"s3": 3, "s4": 2, "s5": 1}
expect_pheno_distances = {"s3": 1, "s4": 2, "s5": 3}
assert got_geno_distances == expect_geno_distances
assert got_pheno_distances == expect_pheno_distances
(
got_geno_distances,
got_pheno_distances,
) = sample_neighbours_finding._geno_and_pheno_distances_for_one_sample(
genos, phenos, pheno_compare, "s1", top_n_genos=3)
expect_geno_distances = {"s4": 2, "s5": 1}
expect_pheno_distances = {"s4": 2, "s5": 3}
assert got_geno_distances == expect_geno_distances
assert got_pheno_distances == expect_pheno_distances
def test_init_in_particular_sanity_check_constraints():
constraints = {
"d1": {
"must_be_same": True,
"method": "equal",
"params": {}
}
}
phenotype_compare.PhenotypeCompare(constraints)
# Uses an unknown method
constraints = {
"d1": {
"must_be_same": True,
"method": "WRONG",
"params": {}
}
}
with pytest.raises(RuntimeError):
phenotype_compare.PhenotypeCompare(constraints)
# equal method should not have any params
constraints = {
"d1": {
"must_be_same": True,
"method": "equal",
"params": {
"low": 1
}
}
}
with pytest.raises(RuntimeError):
phenotype_compare.PhenotypeCompare(constraints)
# Range method should have low and high params
constraints = {
"d1": {
"must_be_same": True,
"method": "range",
"params": {}
}
}
with pytest.raises(RuntimeError):
phenotype_compare.PhenotypeCompare(constraints)
constraints = {
"d1": {
"must_be_same": True,
"method": "range",
"params": {
"low": 1
}
}
}
with pytest.raises(RuntimeError):
phenotype_compare.PhenotypeCompare(constraints)
constraints = {
"d1": {
"must_be_same": True,
"method": "range",
"params": {
"low": 1,
"high": 2
}
}
}
phenotype_compare.PhenotypeCompare(constraints)
# abs_distance should have max_dist param
constraints = {
"d1": {
"must_be_same": True,
"method": "abs_distance",
"params": {}
}
}
with pytest.raises(RuntimeError):
phenotype_compare.PhenotypeCompare(constraints)
constraints = {
"d1": {
"must_be_same": True,
"method": "abs_distance",
"params": {
"max_dist": 1
},
}
}
phenotype_compare.PhenotypeCompare(constraints)
# percent_distance should have max_percent param
constraints = {
"d1": {
"must_be_same": True,
"method": "percent_distance",
"params": {}
}
}
with pytest.raises(RuntimeError):
phenotype_compare.PhenotypeCompare(constraints)
constraints = {
"d1": {
"must_be_same": True,
"method": "percent_distance",
"params": {
"max_percent": 42
},
}
}
phenotype_compare.PhenotypeCompare(constraints)