Add: mergePoN

README.md

+# README.md
+
+> Author: Qingwang Chen
+>
+> Email: [yaqing.liu@outlook.com](mailto:yaqing.liu@outlook.com)
+> 
+> Last Updates: 07/04/2021
+
+#### Requirements
+
+- choppy
+- Ali-Cloud
+- Linux
+
+#### Panel of Normals
+The Panel of Normals approach uses a set of matched normal samples to determine the baseline level from which to call CNV events. These matched normal samples should be derived from the same library prep and sequencing workflow that was used for the case sample. This allows the algorithm to subtract out system level biases that are not sample specific.
+
+#### Usage
+```

defaults

     "SENTIEON_LICENSE": "192.168.0.55:8990",
     "dbmills_dir": "oss://pgx-reference-data/GRCh38.d1.vd1/",
     "db_mills": "Mills_and_1000G_gold_standard.indels.hg38.vcf",
-    "sentieon_docker": "registry.cn-shanghai.aliyuncs.com/pgx-docker-registry/sentieon-genomics:v2019.11.28",
+    "sentieon_docker": "registry.cn-shanghai.aliyuncs.com/pgx-docker-registry/sentieon-genomics:v2020.10.07",
     "annovar_docker": "registry.cn-shanghai.aliyuncs.com/pgx-docker-registry/annovar:v2018.04",
     "maftools_docker": "registry.cn-shanghai.aliyuncs.com/pgx-docker-registry/r-base:4.0.2",
     "bcftools_docker": "registry.cn-shanghai.aliyuncs.com/pgx-docker-registry/bcftools:v1.9",

inputs

 {
   "{{ project_name }}.sample_id": "{{ sample_id }}",
-  "{{ project_name }}.normal_recaled_bam": "{{ normal_recaled_bam }}",
-  "{{ project_name }}.normal_recaled_bam_index": "{{ normal_recaled_bam_index }}",
+  "{{ project_name }}.normal_recaled_bam_bai": "{{ normal_recaled_bam_bai | tojson }}",
   "{{ project_name }}.fasta": "{{ fasta }}",
   "{{ project_name }}.ref_dir": "{{ ref_dir }}",
   "{{ project_name }}.dbsnp": "{{ dbsnp }}",

samples.json

+{
+	"vcf": [
+    ["oss://choppy-cromwell-result/test-choppy/..._pon.vcf",]
+    ["oss://choppy-cromwell-result/test-choppy/..._pon.vcf",]
+    ["oss://choppy-cromwell-result/test-choppy/..._pon.vcf",]
+    ],
+	"panel_id": "PON_20210407",
+	"sample_id": "PON_20210407"
+}

tasks/PON.wdl

-task PON {
+task PoN {
 
 	String SENTIEON_INSTALL_DIR
 	String SENTIEON_LICENSE
 	String dbsnp
 	File regions
 
-	File normal_recaled_bam
-	File normal_recaled_bam_index
+	File normal_bam
+	File normal_bam_index
 	String docker
 	String cluster_config
 	String disk_size
 		mkdir -p /cromwell_root/tmp/cosmic/
 		cp ${cosmic_dir}/${cosmic_vcf} /cromwell_root/tmp/cosmic/
 		${SENTIEON_INSTALL_DIR}/bin/sentieon util vcfindex /cromwell_root/tmp/cosmic/${cosmic_vcf}
-		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -t $nt --interval ${regions} -r ${ref_dir}/${fasta} -i ${normal_recaled_bam} --algo TNhaplotyper --detect_pon --cosmic /cromwell_root/tmp/cosmic/${cosmic_vcf} --dbsnp ${dbsnp_dir}/${dbsnp} ${sample}_pon.vcf
+
+		# Generating TNhaplotyper PON
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -t $nt --interval ${regions} -r ${ref_dir}/${fasta} -i ${normal_bam} --algo TNhaplotyper --detect_pon --cosmic /cromwell_root/tmp/cosmic/${cosmic_vcf} --dbsnp ${dbsnp_dir}/${dbsnp} ${sample}_TNhaplotyper_pon.vcf
+
+		# Generating TNhaplotyper2 PON
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -t $nt --interval ${regions} -r ${ref_dir}/${fasta} -i ${normal_bam} --algo TNhaplotyper2 --normal_sample ${sample} ${sample}_TNhaplotyper2_pon.vcf
+
+		# Generating TNscope PON
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -t $nt --interval ${regions} -r ${ref_dir}/${fasta} -i ${normal_bam} --algo TNscope --normal_sample ${sample} ${sample}_TNscope_pon.vcf
 	>>>
 	
 	runtime {
 	}
 
 	output {
-		File pon_vcf = "${sample}_pon.vcf"
+		File TNhaplotyper_pon_vcf = "${sample}_TNhaplotyper_pon.vcf"
+		File TNhaplotyper2_pon_vcf = "${sample}_TNhaplotyper2_pon.vcf"
+		File TNscope_pon_vcf = "${sample}_TNscope_pon.vcf"
 	}
 }

tasks/mergePoN.wdl

+task mergePoN {
+
+	Array[File] pon_vcfs
+	String panel_id
+	String docker
+	String cluster_config
+	String disk_size
+
+	command <<<
+		set -o pipefail
+		set -e
+		mkdir -p /cromwell_root/tmp/bcftools
+
+		for i in ${sep=" " pon_vcfs}
+		do
+			bcftools view $i -Oz -o /cromwell_root/tmp/bcftools/$(basename $i.gz)
+			bcftools index /cromwell_root/tmp/bcftools/$(basename $i.gz) -f
+		done
+
+		bcftools merge -m all -f PASS,. --force-samples /cromwell_root/tmp/bcftools/*.vcf.gz | bcftools plugin fill-AN-AC | bcftools filter -i 'SUM(AC)>1' > ${panel_id}_pon.vcf
+	>>>
+
+	runtime {
+		docker: docker
+		cluster: cluster_config
+		systemDisk: "cloud_ssd 40"
+		dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
+	}
+	
+	output {
+		File panel_of_normal_vcf = "${panel_id}_pon.vcf"
+	}
+}

workflow.wdl

-import "./tasks/PON.wdl" as PON
+import "./tasks/PoN.wdl" as PoN
+import "./tasks/mergePoN.wdl" as mergePoN
 	
 workflow {{ project_name }} {
 	String SENTIEON_INSTALL_DIR
 	String dbsnp
 	File regions
 
-	File normal_recaled_bam
-	File normal_recaled_bam_index
+	Array[Array[File]] normal_recaled_bam_bai
 	String sentieon_docker
 	String cluster_config
 	String disk_size
-
-	call PON.PON as PON {
+	
+	scatter (item in normal_recaled_bam_bai){
+		call PoN.PoN as PoN {
 		input:
 		SENTIEON_INSTALL_DIR=SENTIEON_INSTALL_DIR,
 		SENTIEON_LICENSE=SENTIEON_LICENSE,
-		sample=sample_id,
 		fasta=fasta,
 		ref_dir=ref_dir,
 		regions=regions,
-		normal_recaled_bam=normal_recaled_bam,
-		normal_recaled_bam_index=normal_recaled_bam_index,
+		sample=item[0],
+		normal_bam=item[1],
+		normal_bam_index=item[2],
 		cosmic_vcf=cosmic_vcf,
 		cosmic_dir=cosmic_dir,
 		dbsnp=dbsnp,
 		docker=sentieon_docker,
 		disk_size=disk_size,
 		cluster_config=cluster_config
+		}
+	}
+
+	Array[File] TNhaplotyper_pon_vcfs = PON.TNhaplotyper_pon_vcf
+	Array[File] TNhaplotyper2_pon_vcfs = PON.TNhaplotyper2_pon_vcf
+	Array[File] TNscope_pon_vcfs = PON.TNscope_pon_vcf
+
+	call mergePoN.mergePoN as mergePoN_TNhaplotyper {
+		input:
+		pon_vcfs=TNhaplotyper_pon_vcfs,
+		panel_id=panel_id,
+		docker=bcftools_docker,
+		disk_size=disk_size,
+		cluster_config=cluster_config
+	}
+
+	call mergePoN.mergePoN as mergePoN_TNhaplotyper2 {
+		input:
+		pon_vcfs=TNhaplotyper2_pon_vcfs,
+		panel_id=panel_id,
+		docker=bcftools_docker,
+		disk_size=disk_size,
+		cluster_config=cluster_config
+	}
+	
+	call mergePoN.mergePoN as mergePoN_TNscope {
+		input:
+		pon_vcfs=TNscope_pon_vcfs,
+		panel_id=panel_id,
+		docker=bcftools_docker,
+		disk_size=disk_size,
+		cluster_config=cluster_config
 	}
 }