def get_callset_truth_data(
truth_sample: str,
mt: bool = True) -> Union[MatrixTableResource, TableResource]:
"""
Get resources for the truth sample data that is subset from the full callset
If `mt` this will return the truth sample MatrixTable (subset from callset); otherwise it returns the
merged truth sample Table that includes both the truth data and the data from the callset
:param str truth_sample: Name of the truth sample
:param bool mt: Whether path is for a MatrixTable, default is True
:return: Path to callset truth sample MT
:rtype: str
"""
if mt:
return VersionedMatrixTableResource(
CURRENT_RELEASE,
{
release: MatrixTableResource(
f"{get_variant_qc_root(release)}/truth_samples/{truth_sample}.mt"
)
for release in RELEASES
},
)
else:
return VersionedTableResource(
CURRENT_RELEASE,
{
release: TableResource(
f"{get_variant_qc_root(release)}/truth_samples/{truth_sample}.ht"
)
for release in RELEASES
},
)
def hgdp_1kg_subset(dense: bool = False) -> VersionedMatrixTableResource:
"""
Get the HGDP + 1KG subset release MatrixTableResource.
:param dense: If True, return the dense MT; if False, return the sparse MT
:return: MatrixTableResource for specific subset
"""
return VersionedMatrixTableResource(
CURRENT_RELEASE,
{
release: MatrixTableResource(
f"gs://gnomad/release/{release}/mt/gnomad.genomes.v{release}.hgdp_1kg_subset{f'_dense' if dense else '_sparse'}.mt"
)
for release in RELEASES
if release != "3"
},
)
CURRENT_EXOME_RELEASE = ""
CURRENT_GENOME_RELEASE = "3.0"
CURRENT_GENOME_COVERAGE_RELEASE = "3.0.1"
EXOME_RELEASES = []
GENOME_RELEASES = ["3.0"]
GENOME_COVERAGE_RELEASES = GENOME_RELEASES + ["3.0.1"]
DATA_TYPES = ["genomes"]
GENOME_POPS = ["AFR", "AMI", "AMR", "ASJ", "EAS", "FIN", "NFE", "SAS", "OTH"]
gnomad_syndip = VersionedMatrixTableResource(
default_version="3.0",
versions={
"3.0":
MatrixTableResource(
path=
"gs://gnomad-public/truth-sets/hail-0.2/gnomad_v3_syndip.b38.mt")
},
)
na12878 = VersionedMatrixTableResource(
default_version="3.0",
versions={
"3.0":
MatrixTableResource(
path="gs://gnomad-public/truth-sets/hail-0.2/gnomad_v3_na12878.mt")
},
)
GnomadPublicTableResource(
path=
"gs://gnomad-public-requester-pays/resources/context/grch38_context_vep_annotated.v101.ht",
),
},
)
syndip = VersionedMatrixTableResource(
default_version="20180222",
versions={
"20180222":
GnomadPublicMatrixTableResource(
path=
"gs://gnomad-public-requester-pays/resources/grch38/syndip/syndip.b38_20180222.mt",
import_func=hl.import_vcf,
import_args={
"path":
"gs://gnomad-public-requester-pays/resources/grch38/syndip/full.38.20180222.vcf.gz",
"force_bgz": True,
"min_partitions": 100,
"reference_genome": "GRCh38",
},
)
},
)
syndip_hc_intervals = VersionedTableResource(
default_version="20180222",
versions={
"20180222":
GnomadPublicTableResource(
path=
mixed_site=(hl.len(mt.alleles) > 2)
& hl.any(lambda a: hl.is_indel(mt.alleles[0], a), mt.alleles[1:])
& hl.any(lambda a: hl.is_snp(mt.alleles[0], a), mt.alleles[1:]),
)
mt = hl.experimental.sparse_split_multi(mt, filter_changed_loci=True)
return mt
gnomad_v3_genotypes = VersionedMatrixTableResource(
CURRENT_RELEASE,
{
"3":
MatrixTableResource(
"gs://gnomad/raw/hail-0.2/mt/genomes_v3/gnomad_genomes_v3.repartitioned.mt"
),
"3.1":
MatrixTableResource(
"gs://gnomad/raw/genomes/3.1/gnomad_v3.1_sparse_unsplit.repartitioned.mt"
),
},
)
def qc_temp_prefix(version: str = CURRENT_RELEASE) -> str:
"""
Returns path to temporary QC bucket.
:param version: Version of annotation path to return
:return: Path to bucket with temporary QC data
"""
)
kgp_phase_3 = VersionedMatrixTableResource(
default_version="phase_3_split",
versions={
"phase_3_split": GnomadPublicMatrixTableResource(
path="gs://gnomad-public-requester-pays/resources/grch37/kgp/1000Genomes_phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.split.mt",
import_func=hl.import_vcf,
import_args={
"path": "gs://genomics-public-data/1000-genomes-phase-3/vcf-20150220/ALL.chr*.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf",
"force_bgz": True,
"skip_invalid_loci": True,
"min_partitions": 300,
"reference_genome": "GRCh37",
},
),
"phase_3": GnomadPublicMatrixTableResource(
path="gs://gnomad-public-requester-pays/resources/grch37/kgp/1000Genomes_phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.mt",
import_func=hl.import_vcf,
import_args={
"path": "gs://genomics-public-data/1000-genomes-phase-3/vcf-20150220/ALL.chr*.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf",
"force_bgz": True,
"skip_invalid_loci": True,
"min_partitions": 300,
"reference_genome": "GRCh37",
},
),
},
)
kgp = VersionedTableResource(
kin_expr=relatedness_ht.kin,
ibd0_expr=relatedness_ht.ibd0,
ibd1_expr=relatedness_ht.ibd1,
ibd2_expr=relatedness_ht.ibd2,
))
# QC Sites (gnomAD v2 QC sites, lifted over)
gnomad_v2_qc_sites = TableResource(
"gs://gnomad-public/resources/grch38/gnomad_v2_qc_sites_b38.ht")
# Dense MT of samples at QC sites
qc = VersionedMatrixTableResource(
CURRENT_RELEASE, {
release: MatrixTableResource(
f"gs://gnomad/sample_qc/mt/genomes_v{release}/gnomad_v{release}_qc_mt_v2_sites_dense.mt"
)
for release in RELEASES
})
# PC relate PCA scores
pc_relate_pca_scores = VersionedTableResource(
CURRENT_RELEASE, {
release: TableResource(
f"{get_sample_qc_root(release)}/gnomad_v{release}_qc_mt_v2_sites_pc_scores.ht"
)
for release in RELEASES
})
# PC relate results
relatedness = VersionedTableResource(