fixup: clean up exome and genome structs on release table

browser/help/topics/v4-browser-hts.md

@@ -51,10 +51,6 @@ Row fields:
       - `hemizygote_count`: Number of hemizygous alternate individuals in this genetic ancestry group.
       - `homozygote_count`: Number of homozygous alternate individuals in this genetic ancestry group.
     - `non_ukb`: Struct containing variant frequency information from the non-UKB subset. Includes same fields as above struct (`all`).
-  - `faf95`: Struct containing information about the filtered allele frequency (FAF; 95% confidence interval [CI]).
-    - `grpmax`: Genetic ancestry group maximum (grpmax) FAF.
-    - `grpmax_gen_anc`: Genetic ancestry group associated with grpmax FAF.
-  - `faf99`: Struct containing information about the FAF (99% CI). Contains same subfields as above (`faf95`).
   - `fafmax`: Struct containing information about the maximum FAF.
     - `gnomad`: Struct containing information about the fafmax for all of gnomad for the exome data.
       - `faf95_max`: Max FAF value for the (95% CI).

data-pipeline/src/data_pipeline/datasets/gnomad_v4/gnomad_v4_variants.py

@@ -391,6 +391,13 @@ def freq_joint(ds, subset=None, pop=None, sex=None, raw=False):
     return ds
 
 
+def prepare_table_for_release(variants_table_path):
+    ds = hl.read_table(variants_table_path)
+    ds = ds.annotate(exomes=ds.exomes.drop("faf95", "faf99"), genomes=ds.genomes.drop("faf95", "faf99"))
+    ds = ds.select_globals(mane_select_version=ds.globals.mane_transcripts_version)
+    return ds
+
+
 def prepare_gnomad_v4_variants(exome_variants_path: str, genome_variants_path: str, variants_joint_frequency_path: str):
     exome_variants = prepare_gnomad_v4_variants_helper(exome_variants_path, "exome")
     genome_variants = prepare_gnomad_v4_variants_helper(genome_variants_path, "genome")

data-pipeline/src/data_pipeline/pipelines/gnomad_v4_variants.py

@@ -8,6 +8,7 @@
 
 from data_pipeline.datasets.gnomad_v4.gnomad_v4_variants import (
     prepare_gnomad_v4_variants,
+    prepare_table_for_release,
 )
 
 
@@ -102,6 +103,15 @@
     },
 )
 
+pipeline.add_task(
+    name="prepare_table_for_release",
+    task_function=prepare_table_for_release,
+    output_path=f"{output_sub_dir}/gnomad_v4_variants_a4_for_release.ht",
+    inputs={
+        "variants_table_path": pipeline.get_task("annotate_vrs_ids"),
+    },
+)
+
 ###############################################
 # Outputs
 ###############################################