Update: input is FASTQ

defaults

@@ -1,4 +1,8 @@
 {
+  "fastq_1": "",
+  "fastq_2": "",
+  "deduped_bam": "",
+  "deduped_bam_index": "",
   "fasta": "GRCh38.d1.vd1.fa",
   "ref_dir": "oss://pgx-reference-data/GRCh38.d1.vd1/",
   "dbsnp": "dbsnp_146.hg38.vcf",

inputs

@@ -1,5 +1,7 @@
 {
   "{{ project_name }}.sample_id": "{{ sample_id }}",
+  "{{ project_name }}.fastq_1": "{{ fastq_1 }}",
+  "{{ project_name }}.fastq_2": "{{ fastq_2 }}",
   "{{ project_name }}.deduped_bam": "{{ deduped_bam }}",
   "{{ project_name }}.deduped_bam_index": "{{ deduped_bam_index }}",
   "{{ project_name }}.fasta": "{{ fasta }}",

tasks/Dedup.wdl

@@ -0,0 +1,41 @@
+task Dedup {
+
+  String SENTIEON_INSTALL_DIR
+  String SENTIEON_LICENSE
+  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=${SENTIEON_LICENSE}
+    nt=$(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 = "${sample}_dedup_metrics.txt"
+    File deduped_bam = "${sample}.sorted.deduped.bam"
+    File deduped_bam_index = "${sample}.sorted.deduped.bam.bai"
+  }
+}
+
+
+
+
+
+

tasks/mapping.wdl

@@ -0,0 +1,35 @@
+task mapping {
+
+  File ref_dir
+  String fasta
+  File fastq_1
+  File fastq_2
+
+  String SENTIEON_INSTALL_DIR
+  String SENTIEON_LICENSE
+  String group
+  String sample
+  String pl
+  String docker
+  String cluster_config
+  String disk_size
+
+  command <<<
+    set -o pipefail
+    set -e  
+    export SENTIEON_LICENSE=${SENTIEON_LICENSE}
+    nt=$(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

@@ -1,3 +1,5 @@
+import "./tasks/mapping.wdl" as mapping
+import "./tasks/Dedup.wdl" as Dedup
 import "./tasks/Realigner.wdl" as Realigner
 import "./tasks/BQSR.wdl" as BQSR
 import "./tasks/PoN.wdl" as PoN
@@ -5,10 +7,13 @@ import "./tasks/PoN.wdl" as PoN
 workflow {{ project_name }} {
   String SENTIEON_INSTALL_DIR
   String SENTIEON_LICENSE
+
   String sample_id
+  File? fastq_1
+  File? fastq_2
+  File? deduped_bam
+  File? deduped_bam_index
 
-  File deduped_bam
-  File deduped_bam_index
   File regions
   File ref_dir
   String fasta
@@ -21,50 +26,84 @@ workflow {{ project_name }} {
   String cluster_config
   String disk_size
   
-  
-  call Realigner.Realigner as Realigner {
-    input:
-    SENTIEON_INSTALL_DIR=SENTIEON_INSTALL_DIR,
-    SENTIEON_LICENSE=SENTIEON_LICENSE,
-    fasta=fasta,
-    ref_dir=ref_dir,
-    deduped_bam=deduped_bam,
-    deduped_bam_index=deduped_bam_index,
-    db_mills=db_mills,
-    dbmills_dir=dbmills_dir,
-    docker=sentieon_docker,
-    disk_size=disk_size,
-    cluster_config=cluster_config
-  }
+  if (fastq_1!= "") {
+    call mapping.mapping as mapping {
+      input: 
+      group=sample_id,
+      sample=sample_id,
+      fastq_1=fastq_1,
+      fastq_2=fastq_2,
+      SENTIEON_INSTALL_DIR=SENTIEON_INSTALL_DIR,
+      SENTIEON_LICENSE=SENTIEON_LICENSE,
+      pl="ILLUMINAL",
+      fasta=fasta,
+      ref_dir=ref_dir,
+      docker=sentieon_docker,
+      disk_size=disk_size,
+      cluster_config=cluster_config
+    }
 
-  call BQSR.BQSR as BQSR {
-    input:
-    SENTIEON_INSTALL_DIR=SENTIEON_INSTALL_DIR,
-    SENTIEON_LICENSE=SENTIEON_LICENSE,
-    fasta=fasta,
-    ref_dir=ref_dir,
-    realigned_bam=Realigner.realigner_bam,
-    realigned_bam_index=Realigner.realigner_bam_index,
-    db_mills=db_mills,
-    dbmills_dir=dbmills_dir,
-    dbsnp=dbsnp,
-    dbsnp_dir=dbsnp_dir,
-    docker=sentieon_docker,
-    disk_size=disk_size,
-    cluster_config=cluster_config
+    call Dedup.Dedup as Dedup {
+      input:
+      SENTIEON_INSTALL_DIR=SENTIEON_INSTALL_DIR,
+      SENTIEON_LICENSE=SENTIEON_LICENSE,
+      sorted_bam=mapping.sorted_bam,
+      sorted_bam_index=mapping.sorted_bam_index,
+      sample=sample_id,
+      docker=sentieon_docker,
+      disk_size=disk_size,
+      cluster_config=cluster_config
+    }
+
+    File deduped_bam=Dedup.deduped_bam
+    File deduped_bam_index=Dedup.deduped_bam_index
   }
 
-  call PoN.PoN as PoN {
-    input:
-    SENTIEON_INSTALL_DIR=SENTIEON_INSTALL_DIR,
-    SENTIEON_LICENSE=SENTIEON_LICENSE,
-    fasta=fasta,
-    ref_dir=ref_dir,
-    regions=regions,
-    normal_bam=BQSR.recaled_bam,
-    normal_bam_index=BQSR.recaled_bam_index,
-    docker=sentieon_docker,
-    disk_size=disk_size,
-    cluster_config=cluster_config
+  if (deduped_bam!= "") {
+    call Realigner.Realigner as Realigner {
+      input:
+      SENTIEON_INSTALL_DIR=SENTIEON_INSTALL_DIR,
+      SENTIEON_LICENSE=SENTIEON_LICENSE,
+      fasta=fasta,
+      ref_dir=ref_dir,
+      deduped_bam=deduped_bam,
+      deduped_bam_index=deduped_bam_index,
+      db_mills=db_mills,
+      dbmills_dir=dbmills_dir,
+      docker=sentieon_docker,
+      disk_size=disk_size,
+      cluster_config=cluster_config
+    }
+
+    call BQSR.BQSR as BQSR {
+      input:
+      SENTIEON_INSTALL_DIR=SENTIEON_INSTALL_DIR,
+      SENTIEON_LICENSE=SENTIEON_LICENSE,
+      fasta=fasta,
+      ref_dir=ref_dir,
+      realigned_bam=Realigner.realigner_bam,
+      realigned_bam_index=Realigner.realigner_bam_index,
+      db_mills=db_mills,
+      dbmills_dir=dbmills_dir,
+      dbsnp=dbsnp,
+      dbsnp_dir=dbsnp_dir,
+      docker=sentieon_docker,
+      disk_size=disk_size,
+      cluster_config=cluster_config
+    }
+
+    call PoN.PoN as PoN {
+      input:
+      SENTIEON_INSTALL_DIR=SENTIEON_INSTALL_DIR,
+      SENTIEON_LICENSE=SENTIEON_LICENSE,
+      fasta=fasta,
+      ref_dir=ref_dir,
+      regions=regions,
+      normal_bam=BQSR.recaled_bam,
+      normal_bam_index=BQSR.recaled_bam_index,
+      docker=sentieon_docker,
+      disk_size=disk_size,
+      cluster_config=cluster_config
+    }
   }
 }