Adding use_allele_specific_annotation arg and fixing task with empty …

…input in JointVcfFiltering WDL (#8027) * Small changes to JointVCFFiltering WDL * making default for use_allele_specific_annotations * addressing comments
broadinstitute · Oct 17, 2022 · c95945c · c95945c
1 parent 78fdd8d
commit c95945c
Show file tree

Hide file tree

Showing 2 changed files with 81 additions and 0 deletions.
diff --git a/scripts/vcf_site_level_filtering_cromwell_tests/vcf_site_level_filtering_travis.json b/scripts/vcf_site_level_filtering_cromwell_tests/vcf_site_level_filtering_travis.json
@@ -9,5 +9,10 @@
   "JointVcfFiltering.basename": "test_10_samples",
   "JointVcfFiltering.snp_annotations": "-A ReadPosRankSum -A FS -A SOR -A QD -A AVERAGE_TREE_SCORE -A AVERAGE_ASSEMBLED_HAPS -A AVERAGE_FILTERED_HAPS",
   "JointVcfFiltering.indel_annotations": "-A MQRankSum -A ReadPosRankSum -A FS -A SOR -A QD -A AVERAGE_TREE_SCORE",
+<<<<<<< HEAD
   "JointVcfFiltering.model_backend": "PYTHON_IFOREST"
+=======
+  "JointVcfFiltering.model_backend": "PYTHON_IFOREST",
+  "JointVcfFiltering.use_allele_specific_annotations": false
+>>>>>>> 99a985c7b (Adding use_allele_specific_annotation arg and fixing task with empty input in JointVcfFiltering WDL (#8027))
 }
diff --git a/scripts/vcf_site_level_filtering_wdl/JointVcfFiltering.wdl b/scripts/vcf_site_level_filtering_wdl/JointVcfFiltering.wdl
@@ -27,6 +27,11 @@ workflow JointVcfFiltering {
 		String indel_annotations
 		File? gatk_override
 
+<<<<<<< HEAD
+=======
+		Boolean use_allele_specific_annotations
+
+>>>>>>> 99a985c7b (Adding use_allele_specific_annotation arg and fixing task with empty input in JointVcfFiltering WDL (#8027))
 		String snp_resource_args = "--resource:hapmap,training=true,calibration=true gs://gcp-public-data--broad-references/hg38/v0/hapmap_3.3.hg38.vcf.gz --resource:omni,training=true,calibration=true gs://gcp-public-data--broad-references/hg38/v0/1000G_omni2.5.hg38.vcf.gz --resource:1000G,training=true,calibration=false gs://gcp-public-data--broad-references/hg38/v0/1000G_phase1.snps.high_confidence.hg38.vcf.gz"
 		String indel_resource_args = "--resource:mills,training=true,calibration=true gs://gcp-public-data--broad-references/hg38/v0/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz"
 	}
@@ -46,6 +51,10 @@ workflow JointVcfFiltering {
 			resource_args = snp_resource_args,
 			basename = basename,
 			interval_list = extract_interval_list,
+<<<<<<< HEAD
+=======
+			use_allele_specific_annotations = use_allele_specific_annotations,
+>>>>>>> 99a985c7b (Adding use_allele_specific_annotation arg and fixing task with empty input in JointVcfFiltering WDL (#8027))
 			gatk_override = gatk_override,
 			gatk_docker = gatk_docker
 	}
@@ -59,6 +68,10 @@ workflow JointVcfFiltering {
 			resource_args = indel_resource_args,
 			basename = basename,
 			interval_list = extract_interval_list,
+<<<<<<< HEAD
+=======
+			use_allele_specific_annotations = use_allele_specific_annotations,
+>>>>>>> 99a985c7b (Adding use_allele_specific_annotation arg and fixing task with empty input in JointVcfFiltering WDL (#8027))
 			gatk_override = gatk_override,
 			gatk_docker = gatk_docker
 	}
@@ -102,6 +115,10 @@ workflow JointVcfFiltering {
 				interval_list = score_interval_list,
 				model_files = TrainVariantAnnotationModelSNPs.outputs,
 				resource_args = snp_resource_args,
+<<<<<<< HEAD
+=======
+				use_allele_specific_annotations = use_allele_specific_annotations,
+>>>>>>> 99a985c7b (Adding use_allele_specific_annotation arg and fixing task with empty input in JointVcfFiltering WDL (#8027))
 				gatk_override = gatk_override,
 				gatk_docker = gatk_docker
 		}
@@ -120,6 +137,7 @@ workflow JointVcfFiltering {
 				interval_list = score_interval_list,
 				model_files = TrainVariantAnnotationModelINDELs.outputs,
 				resource_args = indel_resource_args,
+<<<<<<< HEAD
 				gatk_override = gatk_override,
 				gatk_docker = gatk_docker
 		}
@@ -128,6 +146,18 @@ workflow JointVcfFiltering {
 	output {
 		Array[File] variant_filtered_vcf = ScoreVariantAnnotationsINDELs.output_vcf
 		Array[File] variant_filtered_vcf_index = ScoreVariantAnnotationsINDELs.output_vcf_index
+=======
+				use_allele_specific_annotations = use_allele_specific_annotations,
+				gatk_override = gatk_override,
+				gatk_docker = gatk_docker
+		}
+
+	}
+
+	output {
+		Array[File] variant_scored_vcf = ScoreVariantAnnotationsINDELs.output_vcf
+		Array[File] variant_scored_vcf_index = ScoreVariantAnnotationsINDELs.output_vcf_index
+>>>>>>> 99a985c7b (Adding use_allele_specific_annotation arg and fixing task with empty input in JointVcfFiltering WDL (#8027))
 	}
 
 }
@@ -143,6 +173,10 @@ task ExtractVariantAnnotations {
 		String annotations
 		String resource_args
 		File? interval_list
+<<<<<<< HEAD
+=======
+		Boolean use_allele_specific_annotations
+>>>>>>> 99a985c7b (Adding use_allele_specific_annotation arg and fixing task with empty input in JointVcfFiltering WDL (#8027))
 
 		Int memory_mb = 14000
 		Int command_mem = memory_mb - 1000
@@ -157,6 +191,10 @@ task ExtractVariantAnnotations {
 			-V ~{input_vcf} \
 			-O ~{basename}.~{mode} \
 			~{annotations} \
+<<<<<<< HEAD
+=======
+			~{if use_allele_specific_annotations then "--use-allele-specific-annotations" else ""} \
+>>>>>>> 99a985c7b (Adding use_allele_specific_annotation arg and fixing task with empty input in JointVcfFiltering WDL (#8027))
 			~{"-L " + interval_list} \
 			--mode ~{mode} \
 			~{resource_args}
@@ -232,13 +270,18 @@ task ScoreVariantAnnotations {
 		File extracted_training_vcf_index
 		File? interval_list
 		Array[File] model_files
+<<<<<<< HEAD
+=======
+		Boolean use_allele_specific_annotations
+>>>>>>> 99a985c7b (Adding use_allele_specific_annotation arg and fixing task with empty input in JointVcfFiltering WDL (#8027))
 
 		Int memory_mb = 16000
 		Int command_mem = memory_mb - 1000
 	}
 	Int disk_size = ceil(size(vcf, "GB") *2 + 50)
 
 	command {
+<<<<<<< HEAD
 		set -e
 
 		ln -s ~{sep=" . && ln -s " model_files} .
@@ -261,6 +304,39 @@ task ScoreVariantAnnotations {
 	output {
 		File scores = "~{basename}.~{mode}.scores.hdf5"
 		File annots = "~{basename}.~{mode}.annot.hdf5"
+=======
+		zgrep -v '#' ~{vcf} > empty.txt
+		set -e
+
+		if [ -s empty.txt ]; then
+			ln -s ~{sep=" . && ln -s " model_files} .
+
+			export GATK_LOCAL_JAR=~{default="/root/gatk.jar" gatk_override}
+
+			gatk --java-options "-Xmx~{command_mem}m" \
+				ScoreVariantAnnotations \
+				~{"-L " + interval_list} \
+				-V ~{vcf} \
+				-O ~{basename}.~{mode} \
+				--model-backend ~{model_backend} \
+				~{"--python-script " + python_script} \
+				--model-prefix ~{basename} \
+				~{annotations} \
+				~{if use_allele_specific_annotations then "--use-allele-specific-annotations" else ""} \
+				-mode ~{mode} \
+				--resource:extracted,extracted=true ~{extracted_training_vcf} \
+				~{resource_args}
+		else
+			echo "Input VCF was empty so we'll return the same VCF that was input."
+			echo "Scores and annot hdf5 files will not be produced since the input was empty."
+			ln -s ~{vcf} ~{basename}.~{mode}.vcf.gz
+			ln -s ~{vcf_index} ~{basename}.~{mode}.vcf.gz.tbi
+		fi
+	}
+	output {
+		File? scores = "~{basename}.~{mode}.scores.hdf5"
+		File? annots = "~{basename}.~{mode}.annot.hdf5"
+>>>>>>> 99a985c7b (Adding use_allele_specific_annotation arg and fixing task with empty input in JointVcfFiltering WDL (#8027))
 		File output_vcf = "~{basename}.~{mode}.vcf.gz"
 		File output_vcf_index = "~{basename}.~{mode}.vcf.gz.tbi"
 	}