mito dataset changes (#698)

* handle clinvar change interpretation to classification

* fix trailing comma

* fix some test tables

* update dbnsfp_mito too

* ruff format'

Author: jklugherz

v03_pipeline/lib/annotations/enums.py

@@ -204,7 +204,7 @@
     'Likely_pathogenic/Likely_risk_allele',
     'Established_risk_allele',
     'Likely_risk_allele',
-    'Conflicting_interpretations_of_pathogenicity',
+    'Conflicting_classifications_of_pathogenicity',
     'Uncertain_risk_allele',
     'Uncertain_significance/Uncertain_risk_allele',
     'Uncertain_significance',

v03_pipeline/lib/reference_data/clinvar.py

@@ -25,11 +25,11 @@
 ]
 CLINVAR_GOLD_STARS_LOOKUP = hl.dict(
     {
-        'no_interpretation_for_the_single_variant': 0,
+        'no_classification_for_the_single_variant': 0,
         'no_assertion_provided': 0,
         'no_assertion_criteria_provided': 0,
         'criteria_provided,_single_submitter': 1,
-        'criteria_provided,_conflicting_interpretations': 1,
+        'criteria_provided,_conflicting_classifications': 1,
         'criteria_provided,_multiple_submitters,_no_conflicts': 2,
         'reviewed_by_expert_panel': 3,
         'practice_guideline': 4,

v03_pipeline/lib/tasks/reference_data/update_variant_annotations_table_with_updated_reference_dataset_test.py

@@ -335,7 +335,7 @@ def test_update_vat_with_updated_rdc_mito_38(
                         mitomap='gs://seqr-reference-data/GRCh38/mitochondrial/MITOMAP/mitomap-confirmed-mutations-2022-02-04.ht',
                         mitimpact='gs://seqr-reference-data/GRCh38/mitochondrial/MitImpact/MitImpact_db_3.0.7.ht',
                         clinvar_mito='ftp://ftp.ncbi.nlm.nih.gov/pub/clinvar/vcf_GRCh38/clinvar.vcf.gz',
-                        dbnsfp_mito='gs://seqr-reference-data/GRCh38/dbNSFP/v4.2/dbNSFP4.2a_variant.ht',
+                        dbnsfp_mito='gs://seqr-reference-data/GRCh38/dbNSFP/v4.2/dbNSFP4.2a_variant.with_new_scores.ht',
                         high_constraint_region_mito='gs://seqr-reference-data/GRCh38/mitochondrial/Helix high constraint intervals Feb-15-2022.tsv',
                     ),
                     versions=hl.Struct(
@@ -359,10 +359,7 @@ def test_update_vat_with_updated_rdc_mito_38(
                             assertion=CLINVAR_ASSERTIONS,
                         ),
                         dbnsfp_mito=hl.Struct(
-                            SIFT_pred=['D', 'T'],
-                            Polyphen2_HVAR_pred=['D', 'P', 'B'],
                             MutationTaster_pred=['D', 'A', 'N', 'P'],
-                            fathmm_MKL_coding_pred=['D', 'N'],
                         ),
                         high_constraint_region_mito=hl.Struct(),
                         sorted_transcript_consequences=hl.Struct(
@@ -390,13 +387,8 @@ def test_update_vat_with_updated_rdc_mito_38(
                     alleles=['A', 'C'],
                     clinvar_mito=None,
                     dbnsfp_mito=hl.Struct(
-                        REVEL_score=None,
-                        VEST4_score=None,
-                        MutPred_score=None,
-                        SIFT_pred_id=None,
-                        Polyphen2_HVAR_pred_id=None,
-                        MutationTaster_pred_id=None,
-                        fathmm_MKL_coding_pred_id=None,
+                        SIFT_score=None,
+                        MutationTaster_pred='N',
                     ),
                     gnomad_mito=None,
                     helix_mito=hl.Struct(

v03_pipeline/lib/tasks/update_variant_annotations_table_with_new_samples_test.py

@@ -633,7 +633,7 @@ def test_mito_update_vat(self, mock_update_rdc_task: Mock) -> None:
                     paths=hl.Struct(
                         high_constraint_region_mito='gs://seqr-reference-data/GRCh38/mitochondrial/Helix high constraint intervals Feb-15-2022.tsv',
                         clinvar_mito='ftp://ftp.ncbi.nlm.nih.gov/pub/clinvar/vcf_GRCh38/clinvar.vcf.gz',
-                        dbnsfp_mito='gs://seqr-reference-data/GRCh38/dbNSFP/v4.2/dbNSFP4.2a_variant.ht',
+                        dbnsfp_mito='gs://seqr-reference-data/GRCh38/dbNSFP/v4.2/dbNSFP4.2a_variant.with_new_scores.ht',
                         gnomad_mito='gs://gcp-public-data--gnomad/release/3.1/ht/genomes/gnomad.genomes.v3.1.sites.chrM.ht',
                         helix_mito='gs://seqr-reference-data/GRCh38/mitochondrial/Helix/HelixMTdb_20200327.ht',
                         hmtvar='gs://seqr-reference-data/GRCh38/mitochondrial/HmtVar/HmtVar%20Jan.%2010%202022.ht',
@@ -657,10 +657,7 @@ def test_mito_update_vat(self, mock_update_rdc_task: Mock) -> None:
                             pathogenicity=CLINVAR_PATHOGENICITIES,
                         ),
                         dbnsfp_mito=hl.Struct(
-                            SIFT_pred=['D', 'T'],
-                            Polyphen2_HVAR_pred=['D', 'P', 'B'],
                             MutationTaster_pred=['D', 'A', 'N', 'P'],
-                            fathmm_MKL_coding_pred=['D', 'N'],
                         ),
                         gnomad_mito=hl.Struct(),
                         helix_mito=hl.Struct(),

v03_pipeline/var/test/reference_data/test_combined_1.ht/.README.txt.crc

@@ -1,3 +1,3 @@
 This folder comprises a Hail (www.hail.is) native Table or MatrixTable.
   Written with version 0.2.122-be9d88a80695
-  Created at 2024/01/31 15:18:56
+  Created at 2024/02/21 15:21:48