Call Beta from integration

broadinstitute · Jul 13, 2023 · ed8a73d · ed8a73d
1 parent 4388db6
commit ed8a73d
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 15 deletions.
diff --git a/scripts/variantstore/wdl/GvsJointVariantCalling.wdl b/scripts/variantstore/wdl/GvsJointVariantCalling.wdl
@@ -17,6 +17,11 @@ workflow GvsJointVariantCalling {
         String? vcf_files_column_name
         String? vcf_index_files_column_name
         String? sample_set_name ## NOTE: currently we only allow the loading of one sample set at a time
+
+        # This is the most updated snapshot of the code as of July 2, 2023
+        File gatk_override = "gs://gvs_quickstart_storage/jars/gatk-package-4.2.0.0-724-gf478fb2-SNAPSHOT-local.jar"
+        File interval_list = "gs://gcp-public-data--broad-references/hg38/v0/wgs_calling_regions.hg38.noCentromeres.noTelomeres.interval_list"
+        Boolean extract_do_not_filter_override = false
     }
 
     # the call_set_identifier string is used to name many different things throughout this workflow (BQ tables, vcfs etc),
@@ -40,9 +45,6 @@ workflow GvsJointVariantCalling {
       Int SNP_VQSR_CLASSIC_max_gaussians_override = 6
       Int SNP_VQSR_CLASSIC_mem_gb_override = ""
     }
-    # This is the most updated snapshot of the code as of July 2, 2023
-    File gatk_override = "gs://gvs_quickstart_storage/jars/gatk-package-4.2.0.0-724-gf478fb2-SNAPSHOT-local.jar"
-    File interval_list = "gs://gcp-public-data--broad-references/hg38/v0/wgs_calling_regions.hg38.noCentromeres.noTelomeres.interval_list"
 
     File interval_weights_bed = "gs://broad-public-datasets/gvs/weights/gvs_vet_weights_1kb.bed"
 
@@ -56,7 +58,7 @@ workflow GvsJointVariantCalling {
             extract_output_gcs_dir = extract_output_gcs_dir,
             destination_dataset = dataset_name,
             destination_project = project_id,
-            extract_do_not_filter_override = false,
+            extract_do_not_filter_override = extract_do_not_filter_override,
             extract_maxretries_override = extract_maxretries_override,
             extract_output_file_base_name = extract_output_file_base_name,
             extract_preemptible_override = extract_preemptible_override,
@@ -93,5 +95,6 @@ workflow GvsJointVariantCalling {
         Float total_vcfs_size_mb = GvsUnified.total_vcfs_size_mb
         File? sample_name_list = GvsUnified.sample_name_list
         File manifest = GvsUnified.manifest
+        Boolean done = true
     }
 }
diff --git a/scripts/variantstore/wdl/GvsQuickstartVcfIntegration.wdl b/scripts/variantstore/wdl/GvsQuickstartVcfIntegration.wdl
@@ -2,6 +2,7 @@ version 1.0
 
 import "GvsUnified.wdl" as Unified
 import "GvsUtils.wdl" as Utils
+import "GvsJointVariantCalling.wdl" as JointVariantCalling
 
 workflow GvsQuickstartVcfIntegration {
 
@@ -36,15 +37,13 @@ workflow GvsQuickstartVcfIntegration {
             dataset_suffix = dataset_suffix,
     }
 
-    call Unified.GvsUnified {
+    call JointVariantCalling.GvsJointVariantCalling as Beta {
         input:
             call_set_identifier = branch_name,
             dataset_name = CreateDataset.dataset_name,
             project_id = project_id,
             gatk_override = select_first([gatk_override, BuildGATKJar.jar]),
-            filter_set_name = "quickit",
-            use_VQSR_lite = use_VQSR_lite,
-            extract_table_prefix = "quickit",
+            use_classic_VQSR = !use_VQSR_lite,
             # optionally turn off filtering (VQSR Classic is not deterministic)
             # (and the initial version of this integration test does not allow for inexact matching of actual and expected results.)
             extract_do_not_filter_override = extract_do_not_filter_override,
@@ -62,31 +61,31 @@ workflow GvsQuickstartVcfIntegration {
         call AssertIdenticalOutputs {
             input:
                 expected_output_prefix = expected_prefix,
-                actual_vcfs = GvsUnified.output_vcfs,
+                actual_vcfs = Beta.output_vcfs,
         }
 
         call AssertCostIsTrackedAndExpected {
             input:
-                go = GvsUnified.done,
+                go = Beta.done,
                 dataset_name = CreateDataset.dataset_name,
                 project_id = project_id,
                 expected_output_csv = expected_prefix + "cost_observability_expected.csv",
         }
 
         call AssertTableSizesAreExpected {
             input:
-                go = GvsUnified.done,
+                go = Beta.done,
                 dataset_name = CreateDataset.dataset_name,
                 project_id = project_id,
                 expected_output_csv = expected_prefix + "table_sizes_expected.csv",
         }
     }
 
     output {
-        Array[File] output_vcfs = GvsUnified.output_vcfs
-        Array[File] output_vcf_indexes = GvsUnified.output_vcf_indexes
-        Float total_vcfs_size_mb = GvsUnified.total_vcfs_size_mb
-        File manifest = GvsUnified.manifest
+        Array[File] output_vcfs = Beta.output_vcfs
+        Array[File] output_vcf_indexes = Beta.output_vcf_indexes
+        Float total_vcfs_size_mb = Beta.total_vcfs_size_mb
+        File manifest = Beta.manifest
         String dataset_name = CreateDataset.dataset_name
         String filter_set_name = "quickit"
         Boolean done = true