From 4bb5fb748cbceab1f23d3972cd502e94711fc7c9 Mon Sep 17 00:00:00 2001
From: Sarah Wait Zaranek <swz@curii.com>
Date: Wed, 24 Jun 2020 17:10:55 +0000
Subject: [PATCH] Adding labels Arvados-DCO-1.1-Signed-off-by: Sarah Wait
 Zaranek <swz@curii.com> no issue #

---
 .../cwl/helper/bwamem-gatk-report-wf.cwl      | 61 ++++++++++++++++---
 WGS-processing/cwl/helper/getfastq.cwl        | 29 +++++++--
 WGS-processing/cwl/wgs-processing-wf.cwl      |  5 +-
 3 files changed, 78 insertions(+), 17 deletions(-)

diff --git a/WGS-processing/cwl/helper/bwamem-gatk-report-wf.cwl b/WGS-processing/cwl/helper/bwamem-gatk-report-wf.cwl
index dbe5cf0..f1cfad4 100644
--- a/WGS-processing/cwl/helper/bwamem-gatk-report-wf.cwl
+++ b/WGS-processing/cwl/helper/bwamem-gatk-report-wf.cwl
@@ -1,14 +1,23 @@
 cwlVersion: v1.1
 class: Workflow
+label: WGS processing workflow for single sample
 
 requirements:
   - class: SubworkflowFeatureRequirement
 
 inputs:
-  fastq1: File
-  fastq2: File
+  fastq1: 
+    type: File
+    format: edam:format_1930 # FASTQ
+    label: One of set of pair-end FASTQs (R1)
+  fastq2: 
+    type: File
+    format: edam:format_1930 # FASTQ
+    label: One of set of pair-end FASTQs (R2)
   reference:
     type: File
+    format: edam:format_1929 # FASTA
+    label: Reference genome
     secondaryFiles:
       - .amb
       - .ann
@@ -17,30 +26,55 @@ inputs:
       - .sa
       - .fai
       - ^.dict
-  sample: string
+  sample: 
+    type: string
+    label: Sample Name
   knownsites:
     type: File
+    format: edam:format_3016 # VCF
+    label: VCF of known polymorphic sites for BQSR
     secondaryFiles:
       - .tbi   
-  scattercount: string
-  clinvarvcf: File
-  reportfunc: File
-  headhtml: File
-  tailhtml: File
+  scattercount: 
+    type: string
+    label: Desired split for variant calling
+  clinvarvcf:
+    type: File
+    format: edam:format_3016 # VCF
+    label: Reference VCF for ClinVar
+  reportfunc:
+    type: File
+    label: Function used to create HTML report
+  headhtml:
+    type: File
+    format: edam:format_1964 # HTML
+    label: Header for HTML report
+  tailhtml:
+    type: File
+    format: edam:format_1964 # HTML
+    label: Footer for HTML report
 
 outputs:
   qc-html:
     type: File[]
+    label: FASTQ QC report
+    format: edam:format_1964 # HTML
     outputSource: fastqc/out-html
   qc-zip:
     type: File[]
+    label: Zip file of FASTQ QC report and associated data
     outputSource: fastqc/out-zip 
   gvcf:
     type: File
     outputSource: haplotypecaller/gatheredgvcf
+    format: edam:format_3016 # GVCF
+    label: GVCF generated from GATK 
   report:
     type: File  
     outputSource: generate-report/report
+    format: edam:format_1964 # HTML
+    label: ClinVar variant report
+
 steps:
   fastqc:
     run: fastqc.cwl
@@ -91,3 +125,14 @@ steps:
       headhtml: headhtml
       tailhtml: tailhtml
     out: [report] 
+
+s:codeRepository: https://github.com/arvados/arvados-tutorial
+s:license: https://www.gnu.org/licenses/agpl-3.0.en.html
+
+$namespaces:
+ s: https://schema.org/
+ edam: http://edamontology.org/
+
+$schemas:
+ - https://schema.org/version/latest/schema.rdf
+ - http://edamontology.org/EDAM_1.18.owl
diff --git a/WGS-processing/cwl/helper/getfastq.cwl b/WGS-processing/cwl/helper/getfastq.cwl
index c651771..e2aaa88 100644
--- a/WGS-processing/cwl/helper/getfastq.cwl
+++ b/WGS-processing/cwl/helper/getfastq.cwl
@@ -1,23 +1,28 @@
-$namespaces:
-  arv: "http://arvados.org/cwl#"
-  cwltool: "http://commonwl.org/cwltool#"
 class: ExpressionTool
 cwlVersion: v1.1
-label: Create array of gvcfs to process
+label: Find matching FASTQ pairs 
 requirements:
   InlineJavascriptRequirement: {}
+
 inputs:
   fastqdir:
     type: Directory
-    label: Input directory of fastqs
+    label: Input directory of FASTQs
     loadListing: 'shallow_listing' 
+
 outputs:
   fastq1: 
     type: File[]
+    format: edam:format_1930 # FASTQ
+    label: Half set of pair-end FASTQs (R1)
   fastq2:
     type: File[]
+    format: edam:format_1930 # FASTQ
+    label: Half set of pair-end FASTQs (R2)
   sample:
     type: string[]
+    label: Sample Names
+
 expression: |
   ${function compare(a, b) {
     var baseA = a.basename;
@@ -58,3 +63,17 @@ expression: |
  
     return {"fastq1": fastq1, "fastq2": fastq2, "sample": sample};
   }
+
+
+s:codeRepository: https://github.com/arvados/arvados-tutorial
+s:license: https://www.gnu.org/licenses/agpl-3.0.en.html
+
+$namespaces:
+ s: https://schema.org/
+ edam: http://edamontology.org/
+ arv: "http://arvados.org/cwl#"
+ cwltool: "http://commonwl.org/cwltool#"
+
+$schemas:
+ - https://schema.org/version/latest/schema.rdf
+ - http://edamontology.org/EDAM_1.18.owl
diff --git a/WGS-processing/cwl/wgs-processing-wf.cwl b/WGS-processing/cwl/wgs-processing-wf.cwl
index a6218bd..58e38cb 100644
--- a/WGS-processing/cwl/wgs-processing-wf.cwl
+++ b/WGS-processing/cwl/wgs-processing-wf.cwl
@@ -22,9 +22,6 @@ inputs:
       - .sa
       - .fai
       - ^.dict
-  sample: 
-    type: string
-    label: Sample name
   knownsites:
     type: File
     format: edam:format_3016 # VCF
@@ -55,7 +52,7 @@ outputs:
     type: File[]
     outputSource: bwamem-gatk-report/gvcf
     format: edam:format_3016 # GVCF
-    label: GVCFs generated from sets of fastqs
+    label: GVCFs generated from GATK 
     
   report:
     type: File[]  
-- 
2.30.2