Support vars comply with seqr needs

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 3.0.2
+current_version = 3.0.3
 commit = True
 tag = False
 

.github/workflows/clinvar_runner.yaml

@@ -32,5 +32,5 @@ jobs:
           curl --fail --silent --show-error -X POST \
               -H "Authorization: Bearer $TOKEN" \
               -H "Content-Type:application/json" \
-              -d '{"output": "generate_clinvar_${{ steps.date.outputs.date }}", "dataset": "talos", "accessLevel": "full", "repo": "automated-interpretation-pipeline", "commit": "${{ github.sha }}", "cwd": "reanalysis", "script": ["./clinvar_runner.py"], "description": "Generate Latest Clinvar Summaries", "image": "australia-southeast1-docker.pkg.dev/cpg-common/images/cpg_aip:3.0.2", "config": {"workflow": {"sequencing_type": "genome"}, "cohorts": {"talos": {"clinvar_filter": ["victorian clinical genetics services, murdoch childrens research institute"]}}}, "wait": false}' \
+              -d '{"output": "generate_clinvar_${{ steps.date.outputs.date }}", "dataset": "talos", "accessLevel": "full", "repo": "automated-interpretation-pipeline", "commit": "${{ github.sha }}", "cwd": "reanalysis", "script": ["./clinvar_runner.py"], "description": "Generate Latest Clinvar Summaries", "image": "australia-southeast1-docker.pkg.dev/cpg-common/images/cpg_aip:3.0.3", "config": {"workflow": {"sequencing_type": "genome"}, "cohorts": {"talos": {"clinvar_filter": ["victorian clinical genetics services, murdoch childrens research institute"]}}}, "wait": false}' \
               https://server-a2pko7ameq-ts.a.run.app

.github/workflows/docker.yaml

@@ -14,7 +14,7 @@ permissions:
   contents: read
 
 env:
-  VERSION: 3.0.2
+  VERSION: 3.0.3
 
 jobs:
   docker:

helpers/minimise_output_for_seqr.py

@@ -33,13 +33,15 @@ def coord_to_string(coord: dict) -> str:
     return f"{coord['chrom']}-{coord['pos']}-{coord['ref']}-{coord['alt']}"
 
 
-def main(input_file: str, output: str):
+def main(input_file: str, output: str, ext_map: str | None = None):
     """
     reads in the input file, shrinks it, and writes the output file
 
+
     Args:
         input_file (str):
         output (str):
+        ext_map (str): optional mapping of internal to external IDs for seqr
     """
 
     with open(input_file, encoding='utf-8') as f:
@@ -50,18 +52,24 @@ def main(input_file: str, output: str):
             'metadata': {'categories': data.metadata.categories},
         }
     )
+    ext_map_dict = None
+    if ext_map:
+        with open(ext_map, encoding='utf-8') as f:
+            ext_map_dict = json.load(f)
 
     for individual, details in data.results.items():
+        # optionally update to point to Seqr identities
+        if ext_map_dict:
+            individual = ext_map_dict.get(individual, individual)
+
         lil_data.results[individual] = {}
         for variant in details.variants:
             var_data = variant.var_data
-            lil_data.results[individual][
-                var_data.coordinates.string_format
-            ] = MiniVariant(
+            lil_data.results[individual][var_data.info['seqr_link']] = MiniVariant(
                 **{
                     'categories': variant.categories,
-                    'support_vars': variant.support_vars,
-                    'independent': variant.independent,
+                    'support_vars': variant.support_vars
+                    # 'independent': variant.independent,
                 }
             )
 
@@ -73,6 +81,13 @@ def main(input_file: str, output: str):
     parser = ArgumentParser()
     parser.add_argument('input_file', help='the input file to process')
     parser.add_argument('output_file', help='the output file to write to')
+    parser.add_argument(
+        '--external_map',
+        help='mapping of internal to external IDs for seqr',
+        default=None,
+        type=str,
+        required=False,
+    )
     args = parser.parse_args()
 
-    main(input_file=args.input_file, output=args.output_file)
+    main(input_file=args.input_file, output=args.output_file, ext_map=args.external_map)

reanalysis/moi_tests.py

@@ -570,7 +570,7 @@ def run(
                         genotypes=self.get_family_genotypes(
                             variant=principal, sample_id=sample_id
                         ),
-                        support_vars={partner_variant.coordinates.string_format},
+                        support_vars={partner_variant.info['seqr_link']},
                         flags=principal.get_sample_flags(sample_id)
                         | partner_variant.get_sample_flags(sample_id),
                         independent=False,
@@ -1078,7 +1078,8 @@ def run(
                         genotypes=self.get_family_genotypes(
                             variant=principal, sample_id=sample_id
                         ),
-                        support_vars={partner.coordinates.string_format},
+                        # needs to comply with Seqr
+                        support_vars={partner.info['seqr_link']},
                         flags=principal.get_sample_flags(sample_id)
                         | partner.get_sample_flags(sample_id),
                         independent=False,

reanalysis/reanalysis_global.toml

@@ -74,7 +74,7 @@ default_memory = 'highmem'
 
 [images]
 gatk = 'australia-southeast1-docker.pkg.dev/cpg-common/images/gatk:4.2.6.1'
-aip = 'australia-southeast1-docker.pkg.dev/cpg-common/images/cpg_aip:3.0.2'
+aip = 'australia-southeast1-docker.pkg.dev/cpg-common/images/cpg_aip:3.0.3'
 vep_110 = "australia-southeast1-docker.pkg.dev/cpg-common/images/vep_110:release_110.1"
 cpg_workflows = "australia-southeast1-docker.pkg.dev/cpg-common/images/cpg_workflows:latest"
 

reanalysis/version.py

@@ -3,4 +3,4 @@
 """
 
 # Do not edit this file manually
-__version__ = '3.0.2'
+__version__ = '3.0.3'

setup.py

@@ -31,7 +31,7 @@ def read_reqs(filename: str) -> list[str]:
     name='automated-interpretation-pipeline',
     description='CPG Variant Prioritisation',
     long_description=readme,
-    version='3.0.2',
+    version='3.0.3',
     author='Matthew Welland, CPG',
     author_email=(
         '[email protected], '

test/test_moi_tests.py

@@ -255,7 +255,7 @@ def test_recessive_autosomal_comp_het_male_passes(peddy_ped):
         ab_ratios={'male': 0.5},
         depths={'male': 50},
         boolean_categories=['categoryboolean1'],
-        info={'gene_id': 'TEST1', 'categoryboolean1': True},
+        info={'gene_id': 'TEST1', 'categoryboolean1': True, 'seqr_link': 'passing1'},
         transcript_consequences=[],
     )
     passing_variant2 = SmallVariant(
@@ -264,7 +264,7 @@ def test_recessive_autosomal_comp_het_male_passes(peddy_ped):
         ab_ratios={'male': 0.5},
         depths={'male': 50},
         boolean_categories=['categoryboolean1'],
-        info={'gene_id': 'TEST1', 'categoryboolean1': True},
+        info={'gene_id': 'TEST1', 'categoryboolean1': True, 'seqr_link': 'passing2'},
         transcript_consequences=[],
     )
     comp_hets = {'male': {TEST_COORDS.string_format: [passing_variant2]}}
@@ -287,7 +287,7 @@ def test_recessive_autosomal_comp_het_male_passes_partner_flag(peddy_ped):
         ab_ratios={'male': 0.5},
         depths={'male': 50},
         boolean_categories=['categoryboolean1'],
-        info={'gene_id': 'TEST1', 'categoryboolean1': True},
+        info={'gene_id': 'TEST1', 'categoryboolean1': True, 'seqr_link': 'passing1'},
         transcript_consequences=[],
     )
     passing_variant2 = SmallVariant(
@@ -296,7 +296,7 @@ def test_recessive_autosomal_comp_het_male_passes_partner_flag(peddy_ped):
         ab_ratios={'male': 1.0},
         depths={'male': 50},
         boolean_categories=['categoryboolean1'],
-        info={'gene_id': 'TEST1', 'categoryboolean1': True},
+        info={'gene_id': 'TEST1', 'categoryboolean1': True, 'seqr_link': 'passing2'},
         transcript_consequences=[],
     )
     comp_hets = {'male': {TEST_COORDS.string_format: [passing_variant2]}}
@@ -305,6 +305,7 @@ def test_recessive_autosomal_comp_het_male_passes_partner_flag(peddy_ped):
     assert len(results) == 1
     assert results[0].reasons == {'Autosomal Recessive Comp-Het'}
     assert results[0].flags == {'AB Ratio'}
+    assert results[0].support_vars == {'passing2'}
 
 
 def test_recessive_autosomal_comp_het_female_passes(peddy_ped):
@@ -321,7 +322,7 @@ def test_recessive_autosomal_comp_het_female_passes(peddy_ped):
         ab_ratios={'female': 0.5},
         depths={'female': 50},
         boolean_categories=['categoryboolean1'],
-        info={'gene_id': 'TEST1', 'categoryboolean1': True},
+        info={'gene_id': 'TEST1', 'categoryboolean1': True, 'seqr_link': 'passing1'},
         transcript_consequences=[],
     )
     passing_variant2 = SmallVariant(
@@ -330,7 +331,7 @@ def test_recessive_autosomal_comp_het_female_passes(peddy_ped):
         ab_ratios={'female': 0.5},
         depths={'female': 50},
         boolean_categories=['categoryboolean1'],
-        info={'gene_id': 'TEST1', 'categoryboolean1': True},
+        info={'gene_id': 'TEST1', 'categoryboolean1': True, 'seqr_link': 'passing2'},
         transcript_consequences=[],
     )
     comp_hets = {'female': {TEST_COORDS.string_format: [passing_variant2]}}
@@ -339,6 +340,7 @@ def test_recessive_autosomal_comp_het_female_passes(peddy_ped):
     assert len(results) == 1
     assert results[0].reasons == {'Autosomal Recessive Comp-Het'}
     assert not results[0].flags
+    assert results[0].support_vars == {'passing2'}
 
 
 def test_recessive_autosomal_comp_het_fails_no_ch_return(peddy_ped):
@@ -573,6 +575,7 @@ def test_x_recessive_female_het_passes(peddy_ped):
         info={
             'gene_id': 'TEST1',
             'categorysample4': ['female'],
+            'seqr_link': 'passing1',
         },
         transcript_consequences=[],
     )
@@ -585,6 +588,7 @@ def test_x_recessive_female_het_passes(peddy_ped):
         info={
             'gene_id': 'TEST1',
             'categorysample4': ['female'],
+            'seqr_link': 'passing2',
         },
         transcript_consequences=[],
     )
@@ -593,6 +597,7 @@ def test_x_recessive_female_het_passes(peddy_ped):
     results = x_rec.run(passing_variant, comp_het=comp_hets)
     assert len(results) == 1
     assert results[0].reasons == {'X_RecessiveFemaleCompHet'}
+    assert results[0].support_vars == {'passing2'}
 
 
 def test_het_de_novo_passes(peddy_ped):