first commit

inputs

@@ -0,0 +1,14 @@
+{
+  "{{ project_name }}.fasta": "GRCh38.d1.vd1.fa",
+  "{{ project_name }}.ref_dir": "oss://pgx-reference-data/GRCh38.d1.vd1/",
+  "{{ project_name }}.dbsnp": "dbsnp_146.hg38.vcf",
+  "{{ project_name }}.SENTIEON_INSTALL_DIR": "/opt/sentieon-genomics",
+  "{{ project_name }}.dbmills_dir": "oss://pgx-reference-data/GRCh38.d1.vd1/",
+  "{{ project_name }}.db_mills": "Mills_and_1000G_gold_standard.indels.hg38.vcf",
+  "{{ project_name }}.cluster_config": "{{ cluster if cluster != '' else 'OnDemand ecs.sn1ne.8xlarge img-ubuntu-vpc' }}",
+  "{{ project_name }}.docker": "registry.cn-shanghai.aliyuncs.com/pgx-docker-registry/sentieon-genomics:v2018.08.01",
+  "{{ project_name }}.dbsnp_dir": "oss://pgx-reference-data/GRCh38.d1.vd1/",
+  "{{ project_name }}.sample": "{{ sample_name }}",
+  "{{ project_name }}.disk_size": 500
+}
+

tasks/BQSR.wdl

@@ -0,0 +1,48 @@
+task BQSR {
+	
+    File ref_dir
+    File dbsnp_dir
+    File dbmills_dir
+	String sample
+	String SENTIEON_INSTALL_DIR
+	String fasta
+	String dbsnp
+	String db_mills
+	File realigned_bam
+	File realigned_bam_index
+	String docker
+	String cluster_config
+	String disk_size
+
+	
+	command <<<
+		set -o pipefail
+		set -e
+		export SENTIEON_LICENSE=192.168.0.55:8990
+		nt=$(nproc)
+
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -r ${ref_dir}/${fasta} -t $nt -i ${realigned_bam} --algo QualCal -k ${dbsnp_dir}/${dbsnp} -k ${dbmills_dir}/${db_mills} ${sample}_recal_data.table
+
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -r ${ref_dir}/${fasta} -t $nt -i ${realigned_bam} -q ${sample}_recal_data.table --algo QualCal -k ${dbsnp_dir}/${dbsnp} -k ${dbmills_dir}/${db_mills} ${sample}_recal_data.table.post --algo ReadWriter ${sample}.sorted.deduped.realigned.recaled.bam
+
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -t $nt --algo QualCal --plot --before ${sample}_recal_data.table --after ${sample}_recal_data.table.post ${sample}_recal_data.csv
+
+		${SENTIEON_INSTALL_DIR}/bin/sentieon plot QualCal -o ${sample}_bqsrreport.pdf ${sample}_recal_data.csv
+
+	>>>
+	runtime {
+		docker:docker
+		cluster: cluster_config
+		systemDisk: "cloud_ssd 40"
+		dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"	
+	}
+
+	output {
+		File recal_table = "${sample}_recal_data.table"
+		File recal_post = "${sample}_recal_data.table.post"
+		File recaled_bam = "${sample}.sorted.deduped.realigned.recaled.bam"
+		File recaled_bam_index = "${sample}.sorted.deduped.realigned.recaled.bam.bai"
+		File recal_csv = "${sample}_recal_data.csv"
+		File bqsrreport_pdf = "${sample}_bqsrreport.pdf"
+	}
+}

tasks/Dedup.wdl

@@ -0,0 +1,40 @@
+task Dedup {
+
+	String SENTIEON_INSTALL_DIR
+	String sample
+
+	File sorted_bam
+	File sorted_bam_index
+	String docker
+	String cluster_config
+	String disk_size
+
+
+	command <<<
+		set -o pipefail
+		set -e
+		export SENTIEON_LICENSE=192.168.0.55:8990
+		nt=$(nproc)
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -t $nt -i ${sorted_bam} --algo LocusCollector --fun score_info ${sample}_score.txt
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -t $nt -i ${sorted_bam} --algo Dedup --rmdup --score_info ${sample}_score.txt --metrics ${sample}_dedup_metrics.txt ${sample}.sorted.deduped.bam	
+	>>>
+	runtime {
+		docker:docker
+    	cluster: cluster_config
+    	systemDisk: "cloud_ssd 40"
+    	dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
+	}
+
+	output {
+		File score = "${sample}_score.txt"
+		File dedup_metrics = "${sample}_dedup_metrics.txt"
+		File Dedup_bam = "${sample}.sorted.deduped.bam"
+		File Dedup_bam_index = "${sample}.sorted.deduped.bam.bai"
+	}
+}
+
+
+
+
+
+

tasks/Haplotyper.wdl

@@ -0,0 +1,38 @@
+task Haplotyper {
+	
+	
+    File ref_dir
+    File dbsnp_dir
+
+	String SENTIEON_INSTALL_DIR
+	String fasta
+	File recaled_bam
+	File recaled_bam_index
+	String dbsnp
+	String sample
+	String docker
+	String cluster_config
+	String disk_size
+
+command <<<
+		set -o pipefail
+		set -e
+		export SENTIEON_LICENSE=192.168.0.55:8990
+		nt=$(nproc)	
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -r ${ref_dir}/${fasta} -t $nt -i ${recaled_bam} --algo Haplotyper -d ${dbsnp_dir}/${dbsnp} ${sample}_hc.vcf
+	>>>
+	
+	runtime {
+		docker:docker
+    	cluster: cluster_config
+    	systemDisk: "cloud_ssd 40"
+    	dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
+	}
+
+	output {
+		File vcf = "${sample}_hc.vcf"
+		File vcf_idx = "${sample}_hc.vcf.idx"
+	}
+}
+
+

tasks/Metrics.wdl

@@ -0,0 +1,56 @@
+task Metrics {
+
+
+    File ref_dir
+	String SENTIEON_INSTALL_DIR
+	String sample
+	String docker
+	String cluster_config
+
+	String fasta
+	File sorted_bam
+	File sorted_bam_index
+	String disk_size
+	
+
+
+	command <<<
+		set -o pipefail
+		set -e
+		export SENTIEON_LICENSE=192.168.0.55:8990
+		nt=$(nproc)
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -r ${ref_dir}/${fasta} -t $nt -i ${sorted_bam} --algo MeanQualityByCycle ${sample}_mq_metrics.txt --algo QualDistribution ${sample}_qd_metrics.txt --algo GCBias --summary ${sample}_gc_summary.txt ${sample}_gc_metrics.txt --algo AlignmentStat ${sample}_aln_metrics.txt --algo InsertSizeMetricAlgo ${sample}_is_metrics.txt --algo CoverageMetrics --omit_base_output ${sample}_coverage_metrics
+
+		${SENTIEON_INSTALL_DIR}/bin/sentieon plot metrics -o ${sample}_metrics_report.pdf gc=${sample}_gc_metrics.txt qd=${sample}_qd_metrics.txt mq=${sample}_mq_metrics.txt isize=${sample}_is_metrics.txt
+	>>>
+	
+	runtime {
+		docker:docker
+    	cluster: cluster_config
+    	systemDisk: "cloud_ssd 40"
+    	dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
+	}
+	output {
+		File qd_metrics = "${sample}_qd_metrics.txt"
+		File qd_metrics_pdf = "${sample}_qd_metrics.pdf"
+		File mq_metrics = "${sample}_mq_metrics.txt"
+		File mq_metrics_pdf = "${sample}_mq_metrics.pdf"
+		File is_metrics = "${sample}_is_metrics.txt"
+		File is_metrics_pdf = "${sample}_is_metrics.pdf"
+		File gc_summary = "${sample}_gc_summary.txt"
+		File gc_metrics = "${sample}_gc_metrics.txt"
+		File gc_metrics_pdf = "${sample}_gc_metrics.pdf"
+		File aln_metrics = "${sample}_aln_metrics.txt"
+		File coverage_metrics_sample_summary = "${sample}_coverage_metrics.sample_summary"
+		File coverage_metrics_sample_statistics = "${sample}_coverage_metrics.sample_statistics"
+		File coverage_metrics_sample_interval_statistics = "${sample}_coverage_metrics.sample_interval_statistics"
+		File coverage_metrics_sample_cumulative_coverage_proportions = "${sample}_coverage_metrics.sample_cumulative_coverage_proportions"
+		File coverage_metrics_sample_cumulative_coverage_counts = "${sample}_coverage_metrics.sample_cumulative_coverage_counts"
+	}
+
+}
+
+
+
+
+

tasks/Realigner.wdl

@@ -0,0 +1,41 @@
+task Realigner {
+
+    File ref_dir
+    File dbmills_dir
+
+	String SENTIEON_INSTALL_DIR
+	String sample
+	String fasta
+
+	File Dedup_bam
+	File Dedup_bam_index
+	String db_mills
+	String docker	
+	String cluster_config
+	String disk_size
+
+
+	command <<<
+	set -o pipefail
+	set -e
+	export SENTIEON_LICENSE=192.168.0.55:8990
+	nt=$(nproc)
+	${SENTIEON_INSTALL_DIR}/bin/sentieon driver -r ${ref_dir}/${fasta} -t $nt -i ${Dedup_bam} --algo Realigner -k ${dbmills_dir}/${db_mills} ${sample}.sorted.deduped.realigned.bam
+	
+	>>>
+
+	runtime {
+		docker:docker
+		cluster: cluster_config
+    	systemDisk: "cloud_ssd 40"
+    	dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
+	}
+
+	output {
+		File realigner_bam = "${sample}.sorted.deduped.realigned.bam"
+		File realigner_bam_index = "${sample}.sorted.deduped.realigned.bam.bai"
+
+	}
+}
+
+

tasks/TNscope.wdl

@@ -0,0 +1,42 @@
+task TNscope {
+
+	File ref_dir
+    File dbsnp_dir
+    
+
+	String SENTIEON_INSTALL_DIR
+	String tumor_name
+	String normal_name
+	String docker
+	String cluster_config
+
+	String fasta
+	File corealigner_bam
+	File corealigner_bam_index
+	String dbsnp
+	String disk_size
+
+	
+	
+	command <<<
+		set -o pipefail
+		set -e
+		export SENTIEON_LICENSE=192.168.0.55:8990
+		nt=$(nproc)
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -r ${ref_dir}/${fasta} -t $nt -i ${corealigner_bam} --algo TNscope --tumor_sample ${tumor_name} --normal_sample ${normal_name} --dbsnp ${dbsnp_dir}/${dbsnp} ${sample}.TNscope.TN.vcf
+		
+	>>>
+
+	runtime {
+		docker:docker
+    	cluster: cluster_config
+    	systemDisk: "cloud_ssd 40"
+    	dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
+	}
+
+	output {
+		File TNscope_vcf= "${sample}.TNscope.TN.vcf"
+		File TNscope_vcf_index = "${sample}.TNscope.TN.vcf.idx"
+	}
+
+}

tasks/TNseq.wdl

@@ -0,0 +1,43 @@
+task TNseq {
+	
+    File ref_dir
+    File dbsnp_dir
+
+	String SENTIEON_INSTALL_DIR
+	String tumor_name
+	String normal_name
+	String docker
+	String cluster_config
+
+	String fasta
+	File corealigner_bam
+	File corealigner_bam_index
+	String dbsnp
+	String disk_size
+
+	
+	
+	command <<<
+		set -o pipefail
+		set -e
+		export SENTIEON_LICENSE=192.168.0.55:8990
+		nt=$(nproc)
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -r ${ref_dir}/${fasta} -t $nt -i ${corealigner_bam} --algo TNhaplotyper --tumor_sample ${tumor_name} --normal_sample ${normal_name} --dbsnp ${dbsnp_dir}/${dbsnp} ${sample}.TNseq.TN.vcf
+		
+	>>>
+
+	runtime {
+		docker:docker
+    	cluster: cluster_config
+    	systemDisk: "cloud_ssd 40"
+    	dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
+	}
+
+	output {
+		File TNseq_vcf= "${sample}.TNseq.TN.vcf"
+		File TNseq_vcf_index = "${sample}.TNseq.TN.vcf.idx"
+	}
+
+}
+
+

tasks/corealigner.wdl

@@ -0,0 +1,44 @@
+task corealigner {
+	
+    File ref_dir
+    File dbsnp_dir
+    File dbmills_dir
+
+	String sample
+	String SENTIEON_INSTALL_DIR
+	String docker
+	String cluster_config
+	String fasta
+
+	String dbsnp
+	String db_mills
+	File tumor_recaled_bam
+	File tumor_recaled_bam_index
+	File normal_recaled_bam
+	File normal_recaled_bam_index
+	String disk_size
+
+
+	command <<<
+		set -o pipefail
+		set -e
+		export SENTIEON_LICENSE=192.168.0.55:8990
+		nt=$(nproc)
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -r ${ref_dir}/${fasta} -t $nt -i ${tumor_recaled_bam} -i ${normal_recaled_bam} --algo Realigner -k ${db_mills} -k ${dbsnp} ${sample}_corealigned.bam				
+	>>>
+	
+	runtime {
+		docker:docker
+    	cluster: cluster_config
+    	systemDisk: "cloud_ssd 40"
+    	dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
+	}
+
+	output {
+		File corealigner_bam = "${sample}_corealigned.bam"
+		File corealigner_bam_index = "${sample}_corealigned.bam.bai"
+	}	
+}
+
+
+

tasks/deduped_Metrics.wdl

@@ -0,0 +1,37 @@
+task deduped_Metrics {
+
+    File ref_dir
+    
+	String SENTIEON_INSTALL_DIR
+	String sample
+	String fasta
+	File Dedup_bam
+	File Dedup_bam_index
+	String docker
+	String cluster_config
+	String disk_size
+
+
+	command <<<
+		set -o pipefail
+		set -e
+		export SENTIEON_LICENSE=192.168.0.55:8990
+		nt=$(nproc)
+		${SENTIEON_INSTALL_DIR}/bin/sentieon driver -r ${ref_dir}/${fasta} -t $nt -i ${Dedup_bam} --algo CoverageMetrics --omit_base_output ${sample}_deduped_coverage_metrics --algo MeanQualityByCycle ${sample}_deduped_mq_metrics.txt --algo QualDistribution ${sample}_deduped_qd_metrics.txt --algo GCBias --summary ${sample}_deduped_gc_summary.txt ${sample}_deduped_gc_metrics.txt --algo AlignmentStat ${sample}_deduped_aln_metrics.txt --algo InsertSizeMetricAlgo ${sample}_deduped_is_metrics.txt
+	>>>
+
+	runtime {
+		docker:docker
+    	cluster: cluster_config
+    	systemDisk: "cloud_ssd 40"
+    	dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/" 
+	}
+
+	output {
+		File deduped_coverage_metrics_sample_summary = "${sample}_deduped_coverage_metrics.sample_summary"
+		File deduped_coverage_metrics_sample_statistics = "${sample}_deduped_coverage_metrics.sample_statistics"
+		File deduped_coverage_metrics_sample_interval_statistics = "${sample}_deduped_coverage_metrics.sample_interval_statistics"
+		File deduped_coverage_metrics_sample_cumulative_coverage_proportions = "${sample}_deduped_coverage_metrics.sample_cumulative_coverage_proportions"
+		File deduped_coverage_metrics_sample_cumulative_coverage_counts = "${sample}_deduped_coverage_metrics.sample_cumulative_coverage_counts"
+	}
+}

tasks/mapping.wdl

@@ -0,0 +1,34 @@
+task mapping {
+
+    File ref_dir
+    String fasta
+	File fastq_1
+	File fastq_2
+
+	String SENTIEON_INSTALL_DIR
+	String group
+	String sample
+	String pl
+	String docker
+	String cluster_config
+	String disk_size
+
+	command <<<
+		set -o pipefail
+		set -e	
+		export SENTIEON_LICENSE=192.168.0.55:8990
+		nt=$(nproc)
+		${SENTIEON_INSTALL_DIR}/bin/bwa mem -M -R "@RG\tID:${group}\tSM:${sample}\tPL:${pl}" -t $nt -K 10000000 ${ref_dir}/${fasta} ${fastq_1} ${fastq_2} | ${SENTIEON_INSTALL_DIR}/bin/sentieon util sort -o ${sample}.sorted.bam -t $nt --sam2bam -i -
+	>>>
+
+	runtime {
+		docker:docker
+    	cluster: cluster_config
+    	systemDisk: "cloud_ssd 40"
+    	dataDisk: "cloud_ssd " + disk_size + " /cromwell_root/"
+	}
+	output {
+		File sorted_bam = "${sample}.sorted.bam"
+		File sorted_bam_index = "${sample}.sorted.bam.bai"
+	}
+}

workflow.wdl

@@ -0,0 +1,36 @@
+import "./tasks/Haplotyper.wdl" as Haplotyper
+
+
+workflow {{ project_name }} {
+
+	File recaled_bam
+	File recaled_bam_index
+
+	String SENTIEON_INSTALL_DIR
+	String sample
+	String docker
+	
+	String fasta
+	File ref_dir
+	File dbmills_dir
+	String db_mills
+	File dbsnp_dir
+	String dbsnp
+	String disk_size
+	String cluster_config
+
+	call Haplotyper.Haplotyper as Haplotyper {
+		input:
+		SENTIEON_INSTALL_DIR=SENTIEON_INSTALL_DIR,
+		fasta=fasta,
+		ref_dir=ref_dir,
+		recaled_bam=recaled_bam,
+		recaled_bam_index=recaled_bam_index,
+		dbsnp=dbsnp,
+		dbsnp_dir=dbsnp_dir,
+		sample=sample,
+		docker=docker,
+		disk_size=disk_size,
+		cluster_config=cluster_config
+	}
+}