gnomad_qc.v5.resources.variant_qc
Script containing variant QC related resources for v5.
Module Functions
Features used by the v5 variant QC models, keyed by variant type. |
|
Path to a single-column TSV listing the GCS path to each truth-sample gVCF (one per line). |
|
|
Return the path to the truth-samples VDS combiner plan (combiner |
|
Get the results of variant QC filtering for a given run. |
|
Get the GCS prefix for an isolation forest run's GATK working/output files. |
Script containing variant QC related resources for v5.
- gnomad_qc.v5.resources.variant_qc.VARIANT_QC_FEATURES = {'indel': ['AS_QD', 'AS_pab_max', 'AS_MQRankSum', 'AS_ReadPosRankSum', 'AS_FS', 'AS_SOR'], 'snv': ['AS_QD', 'AS_pab_max', 'AS_MQRankSum', 'AS_ReadPosRankSum', 'AS_FS', 'AS_SOR', 'AS_MQ']}
Features used by the v5 variant QC models, keyed by variant type.
- gnomad_qc.v5.resources.variant_qc.truth_samples_gvcf_paths = 'gs://fc-11093c2b-590e-424a-91ac-0cc040d562fc/v5.0/variant_qc/genomes/aou/truth_samples/truth_samples_gvcf_paths.tsv'
Path to a single-column TSV listing the GCS path to each truth-sample gVCF (one per line).
The truth samples are 8 Genomes-in-a-Bottle (GiaB) samples sequenced with the same protocol as the AoU v8 data. Their gVCFs are stored here: gs://fc-aou-datasets-controlled/v8/wgs/short_read/snpindel/aux/qc/control_samples/
The sample IDs are intentionally not stored in this repo. The combiner therefore reads the gVCF paths from this manifest by known object paths rather than globbing the bucket.
- gnomad_qc.v5.resources.variant_qc.truth_samples_vds = VariantDatasetResource(path=gs://fc-11093c2b-590e-424a-91ac-0cc040d562fc/v5.0/variant_qc/genomes/aou/truth_samples/truth_samples.vds)
VDS containing 8 GiaB samples.
This resource does not need to be remade for future versions.
- gnomad_qc.v5.resources.variant_qc.get_truth_samples_combiner_plan(test=False)[source]
Return the path to the truth-samples VDS combiner plan (combiner
save_path).The plan lets a failed or interrupted combiner run be resumed. It is written to the writable batch variant QC tree (or a temp path for tests).
Like truth_samples_vds, this resource does not need to be remade for future versions.
- Parameters:
test (
bool) – Whether to return a temporary test path. Default is False.- Return type:
str- Returns:
Path to the combiner plan JSON.
- gnomad_qc.v5.resources.variant_qc.get_variant_qc_result(model_id, test=False, split=True, data_type='genomes', environment='batch')[source]
Get the results of variant QC filtering for a given run.
- Parameters:
model_id (
str) – Model ID of variant QC run to load. Must start withrf_,vqsr_, orif_.test (
bool) – Whether to use a tmp path for variant QC tests.split (
bool) – Whether to return the split or unsplit variant QC result.data_type (
str) – Whether to return ‘exomes’ or ‘genomes’ data. Default is genomes.environment (
str) – Environment to use. Default is “batch”. Must be one of “rwb”, “batch”, or “dataproc”.
- Return type:
VersionedTableResource- Returns:
VersionedTableResource for variant QC results.
- gnomad_qc.v5.resources.variant_qc.get_iforest_run_prefix(model_id, version='5.0', test=False, data_type='genomes', environment='batch')[source]
Get the GCS prefix for an isolation forest run’s GATK working/output files.
The isolation forest workflow writes its GATK intermediate outputs (extracted annotations, trained model, scored VCFs) under this prefix, alongside the model’s result HT from
get_variant_qc_result().- Parameters:
model_id (
str) – Model ID of the isolation forest run. Must start withif_.version (
str) – Version of the variant QC result path to return.test (
bool) – Whether to use a tmp path for variant QC tests.data_type (
str) – Whether to return ‘exomes’ or ‘genomes’ data. Default is genomes.environment (
str) – Environment to use. Default is “batch”. Must be one of “rwb”, “batch”, or “dataproc”.
- Return type:
str- Returns:
GCS prefix (no trailing slash) for the isolation forest run’s files.