def test_1kg_chr22():
out_file = new_temp_file(extension='mt')
sample_names = all_samples[:5]
paths = [os.path.join(resource('gvcfs'), '1kg_chr22', f'{s}.hg38.g.vcf.gz') for s in sample_names]
vc.run_combiner(paths,
out_file=out_file,
tmp_path=Env.hc()._tmpdir,
branch_factor=2,
batch_size=2,
reference_genome='GRCh38')
sample_data = dict()
for sample, path in zip(sample_names, paths):
ht = hl.import_vcf(path, force_bgz=True, reference_genome='GRCh38').localize_entries('entries')
n, n_variant = ht.aggregate((hl.agg.count(), hl.agg.count_where(ht.entries[0].GT.is_non_ref())))
sample_data[sample] = (n, n_variant)
mt = hl.read_matrix_table(out_file)
mt = mt.annotate_cols(n=hl.agg.count(), n_variant=hl.agg.count_where(
mt.LGT.is_non_ref())) # annotate the number of non-missing records
combined_results = hl.tuple([mt.s, mt.n, mt.n_variant]).collect()
assert len(combined_results) == len(sample_names)
for sample, n, n_variant in combined_results:
true_n, true_n_variant = sample_data[sample]
assert n == true_n, sample
assert n_variant == true_n_variant, sample
def test_contig_recoding():
path1 = os.path.join(resource('gvcfs'), 'recoding',
'HG00187.hg38.g.vcf.gz')
path2 = os.path.join(resource('gvcfs'), 'recoding',
'HG00187.hg38.recoded.g.vcf.gz')
out_file_1 = new_temp_file(extension='mt')
out_file_2 = new_temp_file(extension='mt')
vc.run_combiner([path1, path1],
out_file_1,
Env.hc()._tmpdir,
reference_genome='GRCh38',
use_exome_default_intervals=True)
vc.run_combiner([path2, path2],
out_file_2,
Env.hc()._tmpdir,
reference_genome='GRCh38',
contig_recoding={'22': 'chr22'},
use_exome_default_intervals=True)
mt1 = hl.read_matrix_table(out_file_1)
mt2 = hl.read_matrix_table(out_file_2)
assert mt1.count() == mt2.count()
assert mt1._same(mt2)
def full_combiner_chr22(*paths):
with TemporaryDirectory() as tmpdir:
vc_all.run_combiner(list(paths),
out_file=tmpdir,
tmp_path='/tmp',
branch_factor=16,
reference_genome='GRCh38',
overwrite=True)
def test_key_by_locus_alleles():
out_file = new_temp_file(extension='mt')
sample_names = all_samples[:5]
paths = [os.path.join(resource('gvcfs'), '1kg_chr22', f'{s}.hg38.g.vcf.gz') for s in sample_names]
vc.run_combiner(paths,
out_file=out_file,
tmp_path=Env.hc()._tmpdir,
reference_genome='GRCh38',
key_by_locus_and_alleles=True)
mt = hl.read_matrix_table(out_file)
assert(list(mt.row_key) == ['locus', 'alleles'])
mt._force_count_rows()
def test_non_ref_alleles_set_to_missing():
path = os.path.join(resource('gvcfs'), 'non_ref_call.g.vcf.gz')
out_file = new_temp_file(extension='mt')
vc.run_combiner([path, path],
out_file=out_file,
tmp_path=Env.hc()._tmpdir,
branch_factor=2,
batch_size=2,
reference_genome='GRCh38')
mt = hl.read_matrix_table(out_file)
n_alleles = hl.len(mt.alleles)
gt_idx = hl.experimental.lgt_to_gt(mt.LGT, mt.LA).unphased_diploid_gt_index()
assert mt.aggregate_entries(
hl.agg.all(gt_idx < (n_alleles * (n_alleles + 1)) / 2))
def test_sample_override():
out_file = new_temp_file(extension='mt')
sample_names = all_samples[:5]
new_names = [f'S{i}' for i, _ in enumerate(sample_names)]
paths = [
os.path.join(resource('gvcfs'), '1kg_chr22', f'{s}.hg38.g.vcf.gz')
for s in sample_names
]
header_path = paths[0]
vc.run_combiner(paths,
out_file=out_file,
tmp_path=Env.hc()._tmpdir,
reference_genome='GRCh38',
header=header_path,
sample_names=new_names,
key_by_locus_and_alleles=True,
use_exome_default_intervals=True)
mt_cols = hl.read_matrix_table(out_file).key_cols_by().cols()
mt_names = mt_cols.aggregate(hl.agg.collect(mt_cols.s))
assert new_names == mt_names
