From 455992936b39ccc8d0e5c94e76c0af382ad8436c Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Fri, 24 Aug 2018 15:16:24 +0300
Subject: [PATCH 01/16] remove operational parameters from scientific
 configuration

---
 .../TOPMed_RNAseq_pipeline/markduplicates.cwl | 14 ++---
 .../TOPMed_RNAseq_pipeline/rna_seqc.cwl       | 17 ++----
 .../rnaseq_pipeline_fastq.cwl                 | 56 ++++++++-----------
 .../TOPMed_RNAseq_pipeline/rsem.cwl           | 10 ++--
 4 files changed, 39 insertions(+), 58 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl
index c31c9d6..53475b6 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     A CWL wrapper for [run_MarkDuplicates.py](https://github.com/broadinstitute/gtex-pipeline/blob/master/rnaseq/src/run_MarkDuplicates.py)
 
@@ -9,12 +10,12 @@ doc: |
 
 cwlVersion: v1.0
 class: CommandLineTool
-id: "run-MarkDuplicates"
 label: "run-MarkDuplicates"
 baseCommand: ["python3", "-u", "/src/run_MarkDuplicates.py"]
 
 requirements:
-  - class: DockerRequirement
+  InlineJavascriptRequirement: {}
+  DockerRequirement:
     dockerPull: heliumdatacommons/topmed-rnaseq:latest
 
 inputs:
@@ -26,11 +27,10 @@ inputs:
     type: string
     inputBinding:
       position: 2
-  memory:
-    type: int
-    inputBinding:
-      position: 3
-      prefix: --memory
+
+arguments:
+  - prefix: --memory
+    valueFrom: ${runtime.mem / 1024}
 
 outputs:
   bam_file:
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rna_seqc.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rna_seqc.cwl
index 7d09bf9..80412c0 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rna_seqc.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rna_seqc.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     A CWL wrapper for [run_rnaseqc.py](https://github.com/heliumdatacommons/cwl_workflows/blob/master/topmed-workflows/TOPMed_RNAseq_pipeline/src/run_rnaseqc.py) duplicated from [run_rnaseqc.py](https://github.com/broadinstitute/gtex-pipeline/blob/master/rnaseq/src/run_rnaseqc.py) with minor modifications.
 
@@ -9,12 +10,12 @@ doc: |
 
 cwlVersion: v1.0
 class: CommandLineTool
-id: "run-seqc"
 label: "run-seqc"
 # run_rnaseqc.py is not an executable file in the docker container.
 baseCommand: ["python3", "/src/run_rnaseqc.py"]
 
 requirements:
+  InlineJavascriptRequirement: {}
   DockerRequirement:
     dockerPull: heliumdatacommons/topmed-rnaseq:latest
 
@@ -40,16 +41,6 @@ inputs:
     type: string
     inputBinding:
       position: 4
-  java_path:
-    type: string
-    inputBinding:
-      position: 5
-      prefix: --java
-  memory:
-    type: int
-    inputBinding:
-      position: 6
-      prefix: --memory
   rnaseqc_flags:
     type:
       type: array
@@ -70,6 +61,10 @@ inputs:
   #     position: 8
   #     prefix: --gatk_flags
 
+arguments:
+  - prefix: --memory
+    valueFrom: ${runtime.ram / 1024}
+
 outputs:
   gene_rpkm:
     type: File
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
index 589cd2e..bafdfbe 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
@@ -15,12 +15,11 @@ doc: |
 
 cwlVersion: v1.0
 class: Workflow
-id: "TOPMed_RNA-seq"
 label: "TOPMed_RNA-seq"
 
 requirements:
-  - class: SubworkflowFeatureRequirement
-  - class: ResourceRequirement
+  SubworkflowFeatureRequirement: {}
+  ResourceRequirement:
     coresMin: 4
     ramMin: 16
     tmpdirMin: 51200
@@ -32,10 +31,6 @@ inputs:
     type: File[]
   prefix_str:
     type: string
-  threads:
-    type: int
-  memory:
-    type: int
   rsem_ref_dir:
     type: Directory
   max_frag_len:
@@ -53,8 +48,6 @@ inputs:
     secondaryFiles:
       - .fai
       - ^.dict
-  java_path:
-    type: string
   rnaseqc_flags:
     type: string[]
   # gatk_flags:
@@ -64,64 +57,64 @@ inputs:
   #     items: string
 
 outputs:
-  - id: star_output_bam
+  star_output_bam:
     outputSource: run_star/bam_file
     type: File
-  - id: star_output_bam_index
+  star_output_bam_index:
     outputSource: run_star/bam_index
     type: File
-  - id: star_output_transcriptome_bam
+  star_output_transcriptome_bam:
     outputSource: run_star/transcriptome_bam
     type: File
-  - id: star_output_chimeric_junctions
+  star_output_chimeric_junctions:
     outputSource: run_star/chimeric_junctions
     type: File
-  - id: star_output_chimeric_bam_file
+  star_output_chimeric_bam_file:
     outputSource: run_star/chimeric_bam_file
     type: File
-  - id: star_output_chimeric_bam_index
+  star_output_chimeric_bam_index:
     outputSource: run_star/chimeric_bam_index
     type: File
-  - id: star_output_read_counts
+  star_output_read_counts:
     outputSource: run_star/read_counts
     type: File
-  - id: star_output_junctions
+  star_output_junctions:
     outputSource: run_star/junctions
     type: File
-  - id: star_output_junctions_pass1
+  star_output_junctions_pass1:
     outputSource: run_star/junctions_pass1
     type: File
-  - id: star_output_logs
+  star_output_logs:
     outputSource: run_star/logs
     type: File[]
-  - id: markduplicates_output_bam
+  markduplicates_output_bam:
     outputSource: run_markduplicates/bam_file
     type: File
-  - id: markduplicates_output_metrics
+  markduplicates_output_metrics:
     outputSource: run_markduplicates/metrics
     type: File
-  - id: markduplicates_bam_index
+  markduplicates_bam_index:
     outputSource: run_index_markduplicates_bam/bam_index
     type: File
-  - id: rsem_output_gene_results
+  rsem_output_gene_results:
     outputSource: run_rsem/gene_results
     type: File
-  - id: rsem_output_isoforms_results
+  rsem_output_isoforms_results:
     outputSource: run_rsem/isoforms_results
     type: File
-  - id: rna-seqc_output_gene_rpkm
+  rna-seqc_output_gene_rpkm:
     outputSource: run_rna-seqc/gene_rpkm
     type: File
-  - id: rna-seqc_output_gene_counts
+  rna-seqc_output_gene_counts:
     outputSource: run_rna-seqc/gene_counts
     type: File
-  - id: rna-seqc_output_exon_counts
+  rna-seqc_output_exon_counts:
     outputSource: run_rna-seqc/exon_counts
     type: File
-  - id: rna-seqc_output_count_metrics
+  rna-seqc_output_count_metrics:
     outputSource: run_rna-seqc/count_metrics
     type: File
-  - id: rna-seqc_output_count_outputs
+  rna-seqc_output_count_outputs:
     outputSource: run_rna-seqc/count_outputs
     type: File
 
@@ -137,7 +130,6 @@ steps:
       star_index: star_index
       fastqs: fastqs
       prefix_str: prefix_str
-      threads: threads
     out:
       [
         bam_file,
@@ -157,7 +149,6 @@ steps:
     in:
       input_bam: run_star/bam_file
       prefix_str: prefix_str
-      memory: memory
     out:
       [
         bam_file,
@@ -180,7 +171,6 @@ steps:
       estimate_rspd: estimate_rspd
       is_stranded: is_stranded
       paired_end: paired_end
-      threads: threads
     out:
       [
         gene_results,
@@ -194,8 +184,6 @@ steps:
       genes_gtf: genes_gtf
       genome_fasta: genome_fasta
       prefix_str: prefix_str
-      java_path: java_path
-      memory: memory
       rnaseqc_flags: rnaseqc_flags
       # gatk_flags: gatk_flags
     out:
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl
index 5de090a..d06ca83 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl
@@ -7,7 +7,6 @@ doc: |
 
 cwlVersion: v1.0
 class: CommandLineTool
-id: "run-rsem"
 label: "run-rsem"
 baseCommand: /src/run_RSEM.py
 
@@ -50,11 +49,10 @@ inputs:
     inputBinding:
       position: 7
       prefix: --paired_end
-  threads:
-    type: int
-    inputBinding:
-      position: 8
-      prefix: --threads
+
+arguments:
+  - prefix: --threads
+    valueFrom: $(runtime.cores)
 
 outputs:
   gene_results:

From a43e0ff4727b80e12143b2f2ac7a7e0bcae76e72 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Fri, 24 Aug 2018 16:05:01 +0300
Subject: [PATCH 02/16] use a plain STAR container

---
 .../rnaseq_pipeline_fastq.cwl                 |  54 +++++---
 .../TOPMed_RNAseq_pipeline/star.cwl           | 118 ++++++++++++++----
 2 files changed, 131 insertions(+), 41 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
index bafdfbe..b25bf8e 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
@@ -58,11 +58,11 @@ inputs:
 
 outputs:
   star_output_bam:
-    outputSource: run_star/bam_file
-    type: File
-  star_output_bam_index:
-    outputSource: run_star/bam_index
+    outputSource: sort_bam/output_file
     type: File
+  # star_output_bam_index:
+  #   outputSource: run_star/bam_index
+  #   type: File
   star_output_transcriptome_bam:
     outputSource: run_star/transcriptome_bam
     type: File
@@ -70,11 +70,11 @@ outputs:
     outputSource: run_star/chimeric_junctions
     type: File
   star_output_chimeric_bam_file:
-    outputSource: run_star/chimeric_bam_file
-    type: File
-  star_output_chimeric_bam_index:
-    outputSource: run_star/chimeric_bam_index
+    outputSource: sort_chimeras/output_file
     type: File
+  # star_output_chimeric_bam_index:
+  #   outputSource: run_star/chimeric_bam_index
+  #   type: File
   star_output_read_counts:
     outputSource: run_star/read_counts
     type: File
@@ -129,25 +129,51 @@ steps:
     in:
       star_index: star_index
       fastqs: fastqs
-      prefix_str: prefix_str
+      prefix: prefix_str
     out:
       [
-        bam_file,
-        bam_index,
+        bam,
         transcriptome_bam,
         chimeric_junctions,
-        chimeric_bam_file,
-        chimeric_bam_index,
+        chimeric_bam,
         read_counts,
         junctions,
         junctions_pass1,
         logs
       ]
 
+  sort_bam:
+    run: https://dockstore.org:8443/api/ga4gh/v2/tools/quay.io%2Fcancercollaboratory%2Fdockstore-tool-samtools-sort/versions/1.0/plain-CWL/descriptor/%2FDockstore.cwl
+    in:
+      threads:
+        valueFrom: $(runtime.cores)
+      memory:
+        valueFrom: $(runtime.ram)M
+      input:
+        source: run_star/bam
+      output_name:
+        source: prefix_str
+        valueFrom: $(self).Aligned.sortedByCoord.out.bam
+    out: [ output_file ]
+
+  sort_chimeras:
+    run: https://dockstore.org:8443/api/ga4gh/v2/tools/quay.io%2Fcancercollaboratory%2Fdockstore-tool-samtools-sort/versions/1.0/plain-CWL/descriptor/%2FDockstore.cwl
+    in:
+      threads:
+        valueFrom: $(runtime.cores)
+      memory:
+        valueFrom: $(runtime.ram)M
+      input:
+        source: run_star/chimeric_bam
+      output_name:
+        source: prefix_str
+        valueFrom: $(self).Chimeric.out.sorted.bam
+    out: [ output_file ]
+
   run_markduplicates:
     run: markduplicates.cwl
     in:
-      input_bam: run_star/bam_file
+      input_bam: sort_bam/output_file
       prefix_str: prefix_str
     out:
       [
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
index ddddabf..7ffb7de 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
@@ -9,21 +9,18 @@ doc: |
 
 cwlVersion: v1.0
 class: CommandLineTool
-id: "run-star"
 label: "run-star"
-baseCommand: /src/run_STAR.py
+baseCommand: STAR
 
-requirements:
+hints:
   DockerRequirement:
-    dockerPull: heliumdatacommons/topmed-rnaseq:latest
+    dockerPull: quay.io/biocontainers/star:2.5.3a--0
 
 inputs:
   star_index:
     type: Directory
-    default:
-      type: Directory
     inputBinding:
-      position: 1
+      prefix: --genomeDir
   fastqs:
     type:
       type: array
@@ -31,26 +28,89 @@ inputs:
       inputBinding:
         itemSeparator: ","
     inputBinding:
-      position: 2
-  prefix_str:
+      prefix: --readFilesIn
+  prefix:
     type: string
     inputBinding:
-      position: 3
-  threads:
-    type: int?
-    inputBinding:
-      position: 5
-      prefix: --threads
+      prefix: --outFileNamePrefix
+      valueFrom: $(runtime.outdir)/$(self).
+
+arguments:
+  - prefix: --runMode
+    valueFrom: alignReads
+  - prefix: --runThreadN
+    valueFrom: $(runtime.cores)
+  - prefix: --twopassMode
+    valueFrom: Basic
+  - prefix: --outFilterMultimapNmax
+    valueFrom: "20"
+  - prefix: --alignSJoverhangMin
+    valueFrom: "8"
+  - prefix: --alignSJDBoverhangMin
+    valueFrom: "1"
+  - prefix: --outFilterMismatchNmax
+    valueFrom: "999"
+  - prefix: --outFilterMismatchNoverLmax
+    valueFrom: "0.1"
+  - prefix: --alignIntronMin
+    valueFrom: "20"
+  - prefix: --alignIntronMax
+    valueFrom: "1000000"
+  - prefix: --outFilterMismatchNoverLmax
+    valueFrom: "0.1"
+  - prefix: --alignMatesGapMax
+    valueFrom: "1000000"
+  - prefix: --outFilterType
+    valueFrom: BySJout
+  - prefix: --outFilterScoreMinOverLread
+    valueFrom: "0.33"
+  - prefix: --outFilterMatchNminOverLread
+    valueFrom: "0.33"
+  - prefix: --limitSjdbInsertNsj
+    valueFrom: "1200000"
+  - prefix: --readFilesCommand
+    valueFrom: zcat
+  - prefix: --outSAMstrandField
+    valueFrom: introMotif
+  - prefix: --outFilterIntronMotifs
+    valueFrom: None
+  - prefix: --alignSoftClipAtReferenceEnds
+    valueFrom: Yes
+  - prefix: --quantMode
+    valueFrom: "TranscriptomeSAM GeneCounts"
+  - prefix: --outSAMtype
+    valueFrom: "BAM Unsorted"
+  - prefix: --outSAMunmapped
+    valueFrom: Within
+  - prefix: --genomeLoad
+    valueFrom: NoSharedMemory
+  - prefix: --chimSegmentMin
+    valueFrom: "15"
+  - prefix: --chimJunctionOverhangMin
+    valueFrom: "15"
+  - prefix: --chimOutType
+    valueFrom: "WithinBAM SoftClip"
+  - prefix: --chimMainSegmentMultNmax
+    valueFrom: "1"
+  - prefix: --outSAMattributes
+    valueFrom: "NH HI AS nM NM ch"
+  - prefix: --outSAMattrRGline
+    valueFrom: "ID:rg1 SM:sm1"
+
 
 outputs:
-  bam_file:
+  bam:
     type: File
     outputBinding:
-      glob: "*.Aligned.sortedByCoord.out.bam"
-  bam_index:
-    type: File
-    outputBinding:
-      glob: "*.Aligned.sortedByCoord.out.bam.bai"
+      glob: $(inputs.prefix).Aligned.out.bam
+  # bam_file:
+  #   type: File
+  #   outputBinding:
+  #     glob: "*.Aligned.sortedByCoord.out.bam"
+  # bam_index:
+  #   type: File
+  #   outputBinding:
+  #     glob: "*.Aligned.sortedByCoord.out.bam.bai"
   transcriptome_bam:
     type: File
     outputBinding:
@@ -59,14 +119,18 @@ outputs:
     type: File
     outputBinding:
       glob: "*.Chimeric.out.junction"
-  chimeric_bam_file:
-    type: File
-    outputBinding:
-      glob: "*.Chimeric.out.sorted.bam"
-  chimeric_bam_index:
+  chimeric_bam:
     type: File
     outputBinding:
-      glob: "*.Chimeric.out.sorted.bam.bai"
+      glob: $(inputs.prefix).Chimeric.out.sam
+  # chimeric_bam_file:
+  #   type: File
+  #   outputBinding:
+  #     glob: "*.Chimeric.out.sorted.bam"
+  # chimeric_bam_index:
+  #   type: File
+  #   outputBinding:
+  #     glob: "*.Chimeric.out.sorted.bam.bai"
   read_counts:
     type: File
     outputBinding:

From ec9d74a8668150390c293dac742d7067fa47c653 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Fri, 24 Aug 2018 17:09:09 +0300
Subject: [PATCH 03/16] index the sorted bams

---
 .../rnaseq_pipeline_fastq.cwl                 | 24 ++++++++++++++-----
 1 file changed, 18 insertions(+), 6 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
index b25bf8e..b8d7981 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
@@ -60,9 +60,9 @@ outputs:
   star_output_bam:
     outputSource: sort_bam/output_file
     type: File
-  # star_output_bam_index:
-  #   outputSource: run_star/bam_index
-  #   type: File
+  star_output_bam_index:
+    outputSource: index_bam/bam_index
+    type: File
   star_output_transcriptome_bam:
     outputSource: run_star/transcriptome_bam
     type: File
@@ -72,9 +72,9 @@ outputs:
   star_output_chimeric_bam_file:
     outputSource: sort_chimeras/output_file
     type: File
-  # star_output_chimeric_bam_index:
-  #   outputSource: run_star/chimeric_bam_index
-  #   type: File
+  star_output_chimeric_bam_index:
+    outputSource: index_chimeras/bas_index
+    type: File
   star_output_read_counts:
     outputSource: run_star/read_counts
     type: File
@@ -170,6 +170,18 @@ steps:
         valueFrom: $(self).Chimeric.out.sorted.bam
     out: [ output_file ]
 
+  index_bam:
+    run: indexbam.cwl
+    in:
+      input_bam: sort_bam/output_file
+    out: [bam_index]
+
+  index_chimeras:
+    run: indexbam.cwl
+    in:
+      input_bam: sort_chimeras/output_file
+    out: [bam_index]
+
   run_markduplicates:
     run: markduplicates.cwl
     in:

From b6216064c347811d58bc82e0dbb8f934720dfb8c Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Fri, 24 Aug 2018 17:23:05 +0300
Subject: [PATCH 04/16] use a plain biocontainer for picard

---
 .../TOPMed_RNAseq_pipeline/indexbam.cwl       | 15 +++++------
 .../TOPMed_RNAseq_pipeline/markduplicates.cwl | 27 ++++++++++---------
 2 files changed, 22 insertions(+), 20 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl
index beb4be4..b1c55fa 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl
@@ -3,23 +3,22 @@ doc: |
 
 cwlVersion: v1.0
 class: CommandLineTool
-id: "run-index-bam"
 label: "run-index-bam"
-baseCommand: ["samtools", "index"]
+baseCommand: [ samtools, index ]
 
 requirements:
-- class: InlineJavascriptRequirement
-- class: DockerRequirement
-  dockerPull: heliumdatacommons/topmed-rnaseq:latest
-- class: InitialWorkDirRequirement
-  listing:
-    - $(inputs.input_bam)
+  DockerRequirement:
+    dockerPull: heliumdatacommons/topmed-rnaseq:latest
+  InitialWorkDirRequirement:
+    listing:
+      - $(inputs.input_bam)
 
 inputs:
   input_bam:
     type: File
     inputBinding:
       position: 1
+      valueFrom: $(self.basename)
 
 outputs:
   bam_index:
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl
index 53475b6..340e171 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl
@@ -11,33 +11,36 @@ doc: |
 cwlVersion: v1.0
 class: CommandLineTool
 label: "run-MarkDuplicates"
-baseCommand: ["python3", "-u", "/src/run_MarkDuplicates.py"]
+baseCommand: [java, picard.cmdline.PicardCommandLine]
 
-requirements:
-  InlineJavascriptRequirement: {}
+hints:
   DockerRequirement:
-    dockerPull: heliumdatacommons/topmed-rnaseq:latest
+    dockerPull: quay.io/biocontainers/picard:2.9.2--2
 
 inputs:
   input_bam:
     type: File
-    inputBinding:
-      position: 1
   prefix_str:
     type: string
-    inputBinding:
-      position: 2
 
 arguments:
-  - prefix: --memory
-    valueFrom: ${runtime.mem / 1024}
+  - prefix: -Xmx
+    valueFrom: $(runtime.ram)M
+  - picard.cmdline.PicardCommandLine
+  - MarkFuplicates
+  - I=$(inputs.input_bam.path)
+  - O=$(runtime.outdir)/$(inputs.input_bam.nameroot).md.bam
+  - M=$(runtime.outdir)/$(inputs.prefix_str).marked_dup_metrics.txt
+  - ASSUME_SORT_ORDER=coordinate
+  - OPTICAL_DUPLICATE_PIXEL_DISTANCE=100
 
 outputs:
   bam_file:
     type: File
     outputBinding:
-      glob: "*.md.bam"
+      glob: $(runtime.outdir)/$(inputs.input_bam.nameroot).md.bam
   metrics:
     type: File
     outputBinding:
-      glob: "*.marked_dup_metrics.txt"
+      glob: $(runtime.outdir)/$(inputs.prefix_str).marked_dup_metrics.txt
+

From 12f4ee5639120f0726c34f5bc50cb194fde7b111 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Fri, 24 Aug 2018 17:24:52 +0300
Subject: [PATCH 05/16] use a plain biocontainer for samtools

---
 topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl
index b1c55fa..ced26a4 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl
@@ -8,7 +8,7 @@ baseCommand: [ samtools, index ]
 
 requirements:
   DockerRequirement:
-    dockerPull: heliumdatacommons/topmed-rnaseq:latest
+    dockerPull: quay.io/biocontainers/samtools:1.8--4
   InitialWorkDirRequirement:
     listing:
       - $(inputs.input_bam)

From c6b26e53020562e6dd1181a30f831e0ace393479 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Fri, 24 Aug 2018 17:56:22 +0300
Subject: [PATCH 06/16] use simple biocontainer for rsem

---
 .../input-examples/rsem-example.yml           |  8 ++--
 .../TOPMed_RNAseq_pipeline/rsem.cwl           | 42 +++++++++----------
 2 files changed, 24 insertions(+), 26 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rsem-example.yml b/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rsem-example.yml
index fab64df..9398d89 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rsem-example.yml
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rsem-example.yml
@@ -1,13 +1,13 @@
 rsem_ref_dir:
     class: Directory
-    location: /rsem_ref/
+    location: ../test-data/rsem_reference
 transcriptome_bam: {
     class: File,
     path: LC_C13_cRNA.Aligned.toTranscriptome.out.bam
 }
 prefix_str: "LC_C13_cRNA"
 max_frag_len: 1000
-estimate_rspd: "true"
-is_stranded: "true"
-paired_end: "true"
+estimate_rspd: true
+is_stranded: true
+paired_end: true
 threads: 4
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl
index d06ca83..6b2ba1c 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     A CWL wrapper for [run_RSEM.py](https://github.com/broadinstitute/gtex-pipeline/blob/master/rnaseq/src/run_RSEM.py)
 
@@ -8,58 +9,55 @@ doc: |
 cwlVersion: v1.0
 class: CommandLineTool
 label: "run-rsem"
-baseCommand: /src/run_RSEM.py
+baseCommand: rsem-calculate-expression
 
-requirements:
+hints:
   DockerRequirement:
-    dockerPull: heliumdatacommons/topmed-rnaseq:latest
+    dockerPull: quay.io/biocontainers/rsem:1.3.0--boost1.64_3
 
 inputs:
   rsem_ref_dir:
     type: Directory
-    default:
-      type: Directory
     inputBinding:
       position: 1
+      valueFrom: $(self.path)/rsem_reference
   transcriptome_bam:
     type: File
     inputBinding:
-      position: 2
+      prefix: --bam
   prefix_str:
     type: string
-    inputBinding:
-      position: 3
   max_frag_len:
     type: int
     inputBinding:
-      position: 4
-      prefix: --max_frag_len
+      prefix: --fragment-length-max
   estimate_rspd:
-    type: string
+    type: boolean
     inputBinding:
-      position: 5
-      prefix: --estimate_rspd
+      prefix: --estimate-rspd
   is_stranded:
-    type: string
+    type: boolean
     inputBinding:
-      position: 6
-      prefix: --is_stranded
+      prefix: "--forward-prob 0.0"
   paired_end:
-    type: string
+    type: boolean
     inputBinding:
-      position: 7
-      prefix: --paired_end
+      prefix: --paired-end
 
 arguments:
-  - prefix: --threads
+  - prefix: --num-threads
     valueFrom: $(runtime.cores)
+  - prefix: --fragment-length-max
+    valueFrom: "1000"
+  - --no-bam-output
+  - $(inputs.prefix_str).rsem
 
 outputs:
   gene_results:
     type: File
     outputBinding:
-      glob: "*.rsem.genes.results"
+      glob: $(inputs.prefix_str).rsem.genes.results
   isoforms_results:
     type: File
     outputBinding:
-      glob: "*.rsem.isoforms.results"
+      glob: $(inputs.prefix_str).rsem.isoforms.results

From ce0ef69c63d6a68721ae4d40afce089cf78d65d9 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Fri, 24 Aug 2018 17:56:53 +0300
Subject: [PATCH 07/16] add #!/usr/bin/env cwl-runner everywhere

---
 .../TOPMed_RNAseq_pipeline/checker-workflows/check_bams_wf.cwl  | 1 +
 .../TOPMed_RNAseq_pipeline/checker-workflows/check_md5_wf.cwl   | 1 +
 .../checker-workflows/components/calc_md5.cwl                   | 1 +
 .../checker-workflows/components/compare_bams.cwl               | 1 +
 .../checker-workflows/components/read_file.cwl                  | 1 +
 .../checker-workflows/components/string_comparison.cwl          | 1 +
 .../checker-workflows/components/untar_dir.cwl                  | 1 +
 .../checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl     | 1 +
 .../checker-workflows/rnaseq_pipeline_fastq_checker.cwl         | 1 +
 topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl            | 1 +
 .../TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl            | 2 ++
 topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl                | 1 +
 12 files changed, 13 insertions(+)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/check_bams_wf.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/check_bams_wf.cwl
index 77743a6..3b1ab51 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/check_bams_wf.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/check_bams_wf.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     Compare 2 input BAM files and report results.
     Exit 0 if sucess.
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/check_md5_wf.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/check_md5_wf.cwl
index 471900b..e4a00bc 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/check_md5_wf.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/check_md5_wf.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     Calculates the MD5 hash of the input file and compares it to the input MD5 hash.
     If hashes match: Exit 0
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/calc_md5.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/calc_md5.cwl
index 9d2a724..e72cf7c 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/calc_md5.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/calc_md5.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     Calculate the MD5 hash for the input file.
 
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/compare_bams.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/compare_bams.cwl
index c65ba33..afa8ff3 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/compare_bams.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/compare_bams.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     Compare 2 input BAM files using [BamUtil diff](https://genome.sph.umich.edu/wiki/BamUtil:_diff)
 
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/read_file.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/read_file.cwl
index d453a93..c24f2df 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/read_file.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/read_file.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 cwlVersion: v1.0
 class: ExpressionTool
 
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/string_comparison.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/string_comparison.cwl
index 7420cf0..88b8a02 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/string_comparison.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/string_comparison.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 cwlVersion: v1.0
 class: CommandLineTool
 baseCommand: []
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/untar_dir.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/untar_dir.cwl
index d5ad119..e4d50b8 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/untar_dir.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/components/untar_dir.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     Extract all files from archive.tar and filter through gzip
 
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl
index ff3d050..9d77366 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     A workflow to verify the proper execution of [TOPMed RNA-seq Workflow](https://github.com/heliumdatacommons/cwl_workflows/blob/master/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl)
 
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker.cwl
index 53a1989..e08d4da 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     A workflow to verify the proper execution of [TOPMed RNA-seq Workflow](https://github.com/heliumdatacommons/cwl_workflows/blob/master/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl)
 
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl
index ced26a4..1451c81 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/indexbam.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     A wrapper for running `samtools index <bam>`.
 
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
index b8d7981..0eff973 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     TOPMed RNA-seq CWL workflow. Documentation on the workflow can be found [here](https://github.com/heliumdatacommons/cwl_workflows/blob/master/topmed-workflows/TOPMed_RNAseq_pipeline/README.md).
     Example input files: [Dockstore.json](https://github.com/heliumdatacommons/cwl_workflows/blob/master/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/Dockstore.json) and [rnaseq_pipeline_fastq-example.yml](https://github.com/heliumdatacommons/cwl_workflows/blob/master/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rnaseq_pipeline_fastq-example.yml).
@@ -19,6 +20,7 @@ label: "TOPMed_RNA-seq"
 
 requirements:
   SubworkflowFeatureRequirement: {}
+hints:
   ResourceRequirement:
     coresMin: 4
     ramMin: 16
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
index 7ffb7de..6b76c73 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
@@ -1,3 +1,4 @@
+#!/usr/bin/env cwl-runner
 doc: |
     A CWL wrapper for [run_STAR.py](https://github.com/broadinstitute/gtex-pipeline/blob/master/rnaseq/src/run_STAR.py)
 

From c27f6b846428c553ce0c42c33a3cf5ae8653c311 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Fri, 24 Aug 2018 18:00:18 +0300
Subject: [PATCH 08/16] two small errors

---
 .../checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl     | 2 --
 .../TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl            | 2 +-
 2 files changed, 1 insertion(+), 3 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl
index 9d77366..2721831 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl
@@ -73,8 +73,6 @@ inputs:
   #   type: string
   # hash_exon_counts:
   #   type: string
-  hash_count_metrics:
-    type: string
   # hash_count_outputs:
   #   type: string
   checker_star_output_bam:
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
index 0eff973..29e4d28 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
@@ -75,7 +75,7 @@ outputs:
     outputSource: sort_chimeras/output_file
     type: File
   star_output_chimeric_bam_index:
-    outputSource: index_chimeras/bas_index
+    outputSource: index_chimeras/bam_index
     type: File
   star_output_read_counts:
     outputSource: run_star/read_counts

From 04b4ecb0f48d337e2bc9fd5cbe05d0bfee6c107f Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Sat, 25 Aug 2018 08:47:15 -0700
Subject: [PATCH 09/16] switch star container

---
 .../rnaseq_pipeline_fastq-example.yml         | 41 +++++++---------
 .../rnaseq_pipeline_fastq.cwl                 | 11 ++---
 .../TOPMed_RNAseq_pipeline/star.cwl           | 49 ++++++++++++++-----
 3 files changed, 56 insertions(+), 45 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rnaseq_pipeline_fastq-example.yml b/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rnaseq_pipeline_fastq-example.yml
index 967e8e9..bfba969 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rnaseq_pipeline_fastq-example.yml
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rnaseq_pipeline_fastq-example.yml
@@ -1,30 +1,23 @@
 star_index:
     class: Directory
-    location: /star_index/
-fastqs: [
-    {class: File,
-    path: LC_C13_cRNA_sequence_R1.fastq.gz},
-    {class: File,
-    path: LC_C13_cRNA_sequence_R2.fastq.gz}
-]
+    location: ../test-data/star_chr12
+fastqs:
+  - class: File
+    path: ../test-data/sample/LC_C13_cRNA_sequence_R1_sub.fastq.gz
+  - class: File
+    path: ../test-data/sample/LC_C13_cRNA_sequence_R2_sub.fastq.gz
 rsem_ref_dir:
     class: Directory
-    location: /rsem_ref/
+    location: ../test-data/rsem_reference
 max_frag_len: 1000
-estimate_rspd: "true"
-is_stranded: "true"
-paired_end: "true"
-genes_gtf: {
-    class: File,
-    path: gencode.v26.annotation.withTranscriptID.gtf
-}
-genome_fasta: {
-    class: File,
-    path: Homo_sapiens_assembly38_noALT_noHLA_noDecoy_ERCC.fasta
-}
-java_path: /usr/lib/jvm/java-1.7.0-openjdk-amd64/bin/java
-memory: 8
+estimate_rspd: true
+is_stranded: true
+paired_end: true
+genes_gtf:
+  class: File
+  path: ../test-data/gencode.v26.annotation.withTranscriptID.gtf
+genome_fasta:
+  class: File
+  path: ../test-data/Homo_sapiens_assembly38_noALT_noHLA_noDecoy_ERCC.fasta
 rnaseqc_flags: ["noDoC", "strictMode"]
-# gatk_flags: []
-prefix_str: "LC_C13_cRNA"
-threads: 4
+prefix_str: LC_C13_cRNA
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
index 29e4d28..9093504 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
@@ -38,11 +38,11 @@ inputs:
   max_frag_len:
     type: int
   estimate_rspd:
-    type: string
+    type: boolean
   is_stranded:
-    type: string
+    type: boolean
   paired_end:
-    type: string
+    type: boolean
   genes_gtf:
     type: File
   genome_fasta:
@@ -122,11 +122,6 @@ outputs:
 
 steps:
   run_star:
-    requirements:
-      ResourceRequirement:
-        coresMin: 4
-        ramMin: 16
-        tmpdirMin: 51200
     run: star.cwl
     in:
       star_index: star_index
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
index 6b76c73..ade57a2 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
@@ -15,7 +15,23 @@ baseCommand: STAR
 
 hints:
   DockerRequirement:
-    dockerPull: quay.io/biocontainers/star:2.5.3a--0
+    #dockerPull: quay.io/biocontainers/star:2.5.3a--0
+    dockerFile: |
+      FROM biocontainers/biocontainers:debian-stretch-backports
+      MAINTAINER biocontainers <biodocker@gmail.com>
+      LABEL    software="rna-star" \
+          container="rna-star" \
+          about.summary="ultrafast universal RNA-seq aligner" \
+          about.home="https://github.com/alexdobin/STAR/" \
+          software.version="2.6.1adfsg-1-deb" \
+          version="1" \
+          about.copyright="2009-2015 Alexander Dobin <dobin@cshl.edu>" \
+          about.license="GPL-3+" \
+          about.license_file="/usr/share/doc/rna-star/copyright" \
+          extra.binaries="/usr/bin/STAR" \
+          about.tags="biology::nucleic-acids, field::biology, field::biology:bioinformatics,:c++, role::program, use::analysing,:biological-sequence"
+      ENV DEBIAN_FRONTEND noninteractive
+      RUN apt-get update && apt-get install -y rna-star && apt-get clean && apt-get purge && rm -rf /var/lib/apt/lists/* /tmp/*
 
 inputs:
   star_index:
@@ -57,8 +73,6 @@ arguments:
     valueFrom: "20"
   - prefix: --alignIntronMax
     valueFrom: "1000000"
-  - prefix: --outFilterMismatchNoverLmax
-    valueFrom: "0.1"
   - prefix: --alignMatesGapMax
     valueFrom: "1000000"
   - prefix: --outFilterType
@@ -77,10 +91,12 @@ arguments:
     valueFrom: None
   - prefix: --alignSoftClipAtReferenceEnds
     valueFrom: Yes
-  - prefix: --quantMode
-    valueFrom: "TranscriptomeSAM GeneCounts"
-  - prefix: --outSAMtype
-    valueFrom: "BAM Unsorted"
+  - --quantMode
+  - TranscriptomeSAM
+  - GeneCounts
+  - --outSAMtype
+  - BAM
+  - Unsorted
   - prefix: --outSAMunmapped
     valueFrom: Within
   - prefix: --genomeLoad
@@ -89,14 +105,21 @@ arguments:
     valueFrom: "15"
   - prefix: --chimJunctionOverhangMin
     valueFrom: "15"
-  - prefix: --chimOutType
-    valueFrom: "WithinBAM SoftClip"
+  - --chimOutType
+  - WithinBAM
+  - SoftClip
   - prefix: --chimMainSegmentMultNmax
     valueFrom: "1"
-  - prefix: --outSAMattributes
-    valueFrom: "NH HI AS nM NM ch"
-  - prefix: --outSAMattrRGline
-    valueFrom: "ID:rg1 SM:sm1"
+  - --outSAMattributes
+  - NH
+  - HI
+  - AS
+  - nM
+  - NM
+  - ch
+  - --outSAMattrRGline
+  - ID:rg1
+  - SM:sm1
 
 
 outputs:

From 2818383766b8843b198d25d36609f1385ea003a2 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Sat, 25 Aug 2018 09:52:50 -0700
Subject: [PATCH 10/16] use local samtools-sort

---
 .../rnaseq_pipeline_fastq.cwl                 | 25 +++++++------------
 1 file changed, 9 insertions(+), 16 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
index 9093504..3f7bbf0 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rnaseq_pipeline_fastq.cwl
@@ -20,11 +20,12 @@ label: "TOPMed_RNA-seq"
 
 requirements:
   SubworkflowFeatureRequirement: {}
-hints:
-  ResourceRequirement:
-    coresMin: 4
-    ramMin: 16
-    tmpdirMin: 51200
+  StepInputExpressionRequirement: {}
+# hints:
+#   ResourceRequirement:
+#     coresMin: 4
+#     ramMin: 16
+#     #tmpdirMin: 51200
 
 inputs:
   star_index:
@@ -140,12 +141,8 @@ steps:
       ]
 
   sort_bam:
-    run: https://dockstore.org:8443/api/ga4gh/v2/tools/quay.io%2Fcancercollaboratory%2Fdockstore-tool-samtools-sort/versions/1.0/plain-CWL/descriptor/%2FDockstore.cwl
+    run: samtools-sort.cwl
     in:
-      threads:
-        valueFrom: $(runtime.cores)
-      memory:
-        valueFrom: $(runtime.ram)M
       input:
         source: run_star/bam
       output_name:
@@ -154,12 +151,8 @@ steps:
     out: [ output_file ]
 
   sort_chimeras:
-    run: https://dockstore.org:8443/api/ga4gh/v2/tools/quay.io%2Fcancercollaboratory%2Fdockstore-tool-samtools-sort/versions/1.0/plain-CWL/descriptor/%2FDockstore.cwl
+    run: samtools-sort.cwl
     in:
-      threads:
-        valueFrom: $(runtime.cores)
-      memory:
-        valueFrom: $(runtime.ram)M
       input:
         source: run_star/chimeric_bam
       output_name:
@@ -231,7 +224,7 @@ steps:
       ]
 
 $namespaces:
-  s: https://schema.org/
+  s: http://schema.org/
 
 $schemas:
 - http://dublincore.org/2012/06/14/dcterms.rdf

From d86c0cd5f7e730f444bbd7980d8882b5d69152f5 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Sat, 25 Aug 2018 09:53:08 -0700
Subject: [PATCH 11/16] fix STAR container

---
 .../TOPMed_RNAseq_pipeline/star.cwl           | 22 +++++++++----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
index ade57a2..21d8b8e 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/star.cwl
@@ -17,13 +17,13 @@ hints:
   DockerRequirement:
     #dockerPull: quay.io/biocontainers/star:2.5.3a--0
     dockerFile: |
-      FROM biocontainers/biocontainers:debian-stretch-backports
+      FROM debian:buster-slim
       MAINTAINER biocontainers <biodocker@gmail.com>
       LABEL    software="rna-star" \
           container="rna-star" \
           about.summary="ultrafast universal RNA-seq aligner" \
           about.home="https://github.com/alexdobin/STAR/" \
-          software.version="2.6.1adfsg-1-deb" \
+          software.version="2.5.3adfsg-1-deb" \
           version="1" \
           about.copyright="2009-2015 Alexander Dobin <dobin@cshl.edu>" \
           about.license="GPL-3+" \
@@ -31,7 +31,15 @@ hints:
           extra.binaries="/usr/bin/STAR" \
           about.tags="biology::nucleic-acids, field::biology, field::biology:bioinformatics,:c++, role::program, use::analysing,:biological-sequence"
       ENV DEBIAN_FRONTEND noninteractive
-      RUN apt-get update && apt-get install -y rna-star && apt-get clean && apt-get purge && rm -rf /var/lib/apt/lists/* /tmp/*
+      RUN echo \
+        'deb http://snapshot.debian.org/archive/debian/20170906/ buster  main' > /etc/apt/sources.list \
+        && printf "Package: r-*\nPin: origin snapshot.debian.org\nPin-Priority: 990\n" > /etc/apt/preferences.d/snapshot \
+        && apt-get -o Acquire::Check-Valid-Until=false update \
+        && apt-get upgrade -y \
+        && apt-get install -y --no-install-recommends \
+            --allow-unauthenticated rna-star=2.5.3a+dfsg-3 \
+        && apt-get clean && apt-get purge && rm -rf /var/lib/apt/lists/* /tmp/*
+    dockerImageId: star
 
 inputs:
   star_index:
@@ -127,14 +135,6 @@ outputs:
     type: File
     outputBinding:
       glob: $(inputs.prefix).Aligned.out.bam
-  # bam_file:
-  #   type: File
-  #   outputBinding:
-  #     glob: "*.Aligned.sortedByCoord.out.bam"
-  # bam_index:
-  #   type: File
-  #   outputBinding:
-  #     glob: "*.Aligned.sortedByCoord.out.bam.bai"
   transcriptome_bam:
     type: File
     outputBinding:

From b981acb5688f83a883b3b92eabc36ed45a600818 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Sat, 25 Aug 2018 09:53:18 -0700
Subject: [PATCH 12/16] improve rsem

---
 .../TOPMed_RNAseq_pipeline/rsem.cwl           | 20 +++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl
index 6b2ba1c..514dad9 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rsem.cwl
@@ -11,6 +11,12 @@ class: CommandLineTool
 label: "run-rsem"
 baseCommand: rsem-calculate-expression
 
+requirements:
+  InlineJavascriptRequirement: {}
+  InitialWorkDirRequirement:
+    listing:
+      - entry: $(inputs.rsem_ref_dir)
+        writable: true
 hints:
   DockerRequirement:
     dockerPull: quay.io/biocontainers/rsem:1.3.0--boost1.64_3
@@ -19,14 +25,17 @@ inputs:
   rsem_ref_dir:
     type: Directory
     inputBinding:
-      position: 1
-      valueFrom: $(self.path)/rsem_reference
+      position: 2
+      valueFrom: $(self.basename)/rsem_reference
   transcriptome_bam:
     type: File
     inputBinding:
-      prefix: --bam
+      position: 1
   prefix_str:
     type: string
+    inputBinding:
+      position: 3
+      valueFrom: $(self).rsem
   max_frag_len:
     type: int
     inputBinding:
@@ -37,8 +46,6 @@ inputs:
       prefix: --estimate-rspd
   is_stranded:
     type: boolean
-    inputBinding:
-      prefix: "--forward-prob 0.0"
   paired_end:
     type: boolean
     inputBinding:
@@ -50,7 +57,8 @@ arguments:
   - prefix: --fragment-length-max
     valueFrom: "1000"
   - --no-bam-output
-  - $(inputs.prefix_str).rsem
+  - --bam
+  - ${if (inputs.is_stranded) { return ["--forward-prob", "0.0"];} }
 
 outputs:
   gene_results:

From a4a69f69504d7bc77f499f20c01e3a77b6931367 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Sat, 25 Aug 2018 10:24:51 -0700
Subject: [PATCH 13/16] fix markduplicates

---
 .../TOPMed_RNAseq_pipeline/markduplicates.cwl     | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl
index 340e171..60086b5 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/markduplicates.cwl
@@ -11,9 +11,13 @@ doc: |
 cwlVersion: v1.0
 class: CommandLineTool
 label: "run-MarkDuplicates"
-baseCommand: [java, picard.cmdline.PicardCommandLine]
+baseCommand: [java]
 
-hints:
+requirements: # turn back into a hint when the biocontainer has its classpath
+              #   updated
+  EnvVarRequirement:
+    envDef:
+      CLASSPATH: /usr/local/share/picard-2.9.2-2/picard.jar
   DockerRequirement:
     dockerPull: quay.io/biocontainers/picard:2.9.2--2
 
@@ -26,8 +30,9 @@ inputs:
 arguments:
   - prefix: -Xmx
     valueFrom: $(runtime.ram)M
+    separate: false
   - picard.cmdline.PicardCommandLine
-  - MarkFuplicates
+  - MarkDuplicates
   - I=$(inputs.input_bam.path)
   - O=$(runtime.outdir)/$(inputs.input_bam.nameroot).md.bam
   - M=$(runtime.outdir)/$(inputs.prefix_str).marked_dup_metrics.txt
@@ -38,9 +43,9 @@ outputs:
   bam_file:
     type: File
     outputBinding:
-      glob: $(runtime.outdir)/$(inputs.input_bam.nameroot).md.bam
+      glob: $(inputs.input_bam.nameroot).md.bam
   metrics:
     type: File
     outputBinding:
-      glob: $(runtime.outdir)/$(inputs.prefix_str).marked_dup_metrics.txt
+      glob: $(inputs.prefix_str).marked_dup_metrics.txt
 

From 68d12917692cfad47313bf2973de904873f359e4 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Sun, 26 Aug 2018 23:57:10 -0700
Subject: [PATCH 14/16] final validation

---
 .../rnaseq_pipeline_fastq-example.yml         | 32 +++++++++++--------
 .../input-examples/rsem-example.yml           |  2 +-
 .../TOPMed_RNAseq_pipeline/rna_seqc.cwl       |  5 +--
 3 files changed, 22 insertions(+), 17 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rnaseq_pipeline_fastq-example.yml b/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rnaseq_pipeline_fastq-example.yml
index bfba969..e6d6b22 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rnaseq_pipeline_fastq-example.yml
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rnaseq_pipeline_fastq-example.yml
@@ -1,23 +1,27 @@
 star_index:
     class: Directory
-    location: ../test-data/star_chr12
-fastqs:
-  - class: File
-    path: ../test-data/sample/LC_C13_cRNA_sequence_R1_sub.fastq.gz
-  - class: File
-    path: ../test-data/sample/LC_C13_cRNA_sequence_R2_sub.fastq.gz
+    location: /star_index/
+fastqs: [
+    {class: File,
+    path: LC_C13_cRNA_sequence_R1.fastq.gz},
+    {class: File,
+    path: LC_C13_cRNA_sequence_R2.fastq.gz}
+]
 rsem_ref_dir:
     class: Directory
-    location: ../test-data/rsem_reference
+    location: /rsem_ref/
 max_frag_len: 1000
 estimate_rspd: true
 is_stranded: true
 paired_end: true
-genes_gtf:
-  class: File
-  path: ../test-data/gencode.v26.annotation.withTranscriptID.gtf
-genome_fasta:
-  class: File
-  path: ../test-data/Homo_sapiens_assembly38_noALT_noHLA_noDecoy_ERCC.fasta
+genes_gtf: {
+    class: File,
+    path: gencode.v26.annotation.withTranscriptID.gtf
+}
+genome_fasta: {
+    class: File,
+    path: Homo_sapiens_assembly38_noALT_noHLA_noDecoy_ERCC.fasta
+}
 rnaseqc_flags: ["noDoC", "strictMode"]
-prefix_str: LC_C13_cRNA
+# gatk_flags: []
+prefix_str: "LC_C13_cRNA"
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rsem-example.yml b/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rsem-example.yml
index 9398d89..73fa82b 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rsem-example.yml
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/input-examples/rsem-example.yml
@@ -1,6 +1,6 @@
 rsem_ref_dir:
     class: Directory
-    location: ../test-data/rsem_reference
+    location: /rsem_ref/
 transcriptome_bam: {
     class: File,
     path: LC_C13_cRNA.Aligned.toTranscriptome.out.bam
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/rna_seqc.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/rna_seqc.cwl
index 80412c0..206536d 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/rna_seqc.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/rna_seqc.cwl
@@ -63,8 +63,9 @@ inputs:
 
 arguments:
   - prefix: --memory
-    valueFrom: ${runtime.ram / 1024}
-
+    valueFrom: $(runtime.ram / 1024)
+  - prefix: --java
+    valueFrom: /usr/lib/jvm/java-1.7.0-openjdk-amd64/bin/java
 outputs:
   gene_rpkm:
     type: File

From ad671727c9a9bd060f263ed72187ae682b8e158e Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Mon, 27 Aug 2018 00:49:29 -0700
Subject: [PATCH 15/16] add missing samtools sort

---
 .../TOPMed_RNAseq_pipeline/samtools-sort.cwl  | 91 +++++++++++++++++++
 1 file changed, 91 insertions(+)
 create mode 100644 topmed-workflows/TOPMed_RNAseq_pipeline/samtools-sort.cwl

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/samtools-sort.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/samtools-sort.cwl
new file mode 100644
index 0000000..9980e15
--- /dev/null
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/samtools-sort.cwl
@@ -0,0 +1,91 @@
+#!/usr/bin/env cwl-runner
+
+# adapted from https://github.com/common-workflow-language/workflows/tree/master/tools
+
+class: CommandLineTool
+
+$namespaces:
+  dct: http://purl.org/dc/terms/
+$schemas: [ http://dublincore.org/2012/06/14/dcterms.rdf ]
+dct:contributor:
+  foaf:name: Andy Yang
+  foaf:mbox: mailto:ayang@oicr.on.ca
+dct:creator:
+  '@id': http://orcid.org/0000-0001-9102-5681
+  foaf:name: Andrey Kartashov
+  foaf:mbox: mailto:Andrey.Kartashov@cchmc.org
+dct:description: 'Developed at Cincinnati Children’s Hospital Medical Center for the
+  CWL consortium http://commonwl.org/ Original URL: https://github.com/common-workflow-language/workflows'
+cwlVersion: v1.0
+
+
+requirements:
+- class: DockerRequirement
+  dockerPull: quay.io/cancercollaboratory/dockstore-tool-samtools-sort:1.0
+inputs:
+  compression_level:
+    type: int?
+    inputBinding:
+      prefix: -l
+    doc: |
+      Set compression level, from 0 (uncompressed) to 9 (best)
+
+    doc: Set number of sorting and compression threads [1]
+  input:
+    type: File
+    inputBinding:
+      position: 1
+
+    doc: Input bam file.
+  output_name:
+    type: string
+    inputBinding:
+      position: 2
+      prefix: -o
+
+    doc: Desired output filename.
+  sort_by_name:
+    type: boolean?
+    inputBinding:
+      prefix: -n
+
+    doc: Sort by read names (i.e., the QNAME field) rather than by chromosomal coordinates.
+outputs:
+  output_file:
+    type: File
+    outputBinding:
+      glob: $(inputs.output_name)
+
+arguments:
+ - prefix: -@
+   valueFrom: $(runtime.cores)
+ - prefix: -m
+   valueFrom: $(runtime.ram)M
+
+baseCommand: [samtools, sort]
+doc: |
+  Sort alignments by leftmost coordinates, or by read name when -n is used. An appropriate @HD-SO sort order header tag will be added or an existing one updated if necessary.
+
+  Usage: samtools sort [-l level] [-m maxMem] [-o out.bam] [-O format] [-n] -T out.prefix [-@ threads] [in.bam]
+
+  Options:
+  -l INT
+  Set the desired compression level for the final output file, ranging from 0 (uncompressed) or 1 (fastest but minimal compression) to 9 (best compression but slowest to write), similarly to gzip(1)'s compression level setting.
+
+  If -l is not used, the default compression level will apply.
+
+
+  -n
+  Sort by read names (i.e., the QNAME field) rather than by chromosomal coordinates.
+
+  -o FILE
+  Write the final sorted output to FILE, rather than to standard output.
+
+  -O FORMAT
+  Write the final output as sam, bam, or cram.
+
+  By default, samtools tries to select a format based on the -o filename extension; if output is to standard output or no format can be deduced, -O must be used.
+
+  -T PREFIX
+  Write temporary files to PREFIX.nnnn.bam. This option is required.
+

From e578e5f9cb99770c423083da54c6756469e1eba6 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Mon, 27 Aug 2018 10:51:24 +0300
Subject: [PATCH 16/16] --validate clean on CWL in topmed-workflows/

---
 .../rnaseq_pipeline_fastq_checker-tar.cwl     | 17 ++++-------------
 .../rnaseq_pipeline_fastq_checker.cwl         | 19 ++++---------------
 .../TOPMed_RNAseq_pipeline/samtools-sort.cwl  |  5 +----
 3 files changed, 9 insertions(+), 32 deletions(-)

diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl
index 2721831..32626ba 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker-tar.cwl
@@ -21,20 +21,16 @@ inputs:
     type: File[]
   prefix_str:
     type: string
-  threads:
-    type: int
-  memory:
-    type: int
   rsem_ref_dir_tar:
     type: File
   max_frag_len:
     type: int
   estimate_rspd:
-    type: string
+    type: boolean
   is_stranded:
-    type: string
+    type: boolean
   paired_end:
-    type: string
+    type: boolean
   genes_gtf:
     type: File
   genome_fasta:
@@ -42,8 +38,6 @@ inputs:
     secondaryFiles:
       - .fai
       - ^.dict
-  java_path:
-    type: string
   rnaseqc_flags:
     type: string[]
   # gatk_flags:
@@ -156,8 +150,6 @@ steps:
       star_index: untar_star_index/untarred_dir
       fastqs: fastqs
       prefix_str: prefix_str
-      threads: threads
-      memory: memory
       rsem_ref_dir: untar_rsem_reference/untarred_dir
       max_frag_len: max_frag_len
       estimate_rspd: estimate_rspd
@@ -165,7 +157,6 @@ steps:
       paired_end: paired_end
       genes_gtf: genes_gtf
       genome_fasta: genome_fasta
-      java_path: java_path
       rnaseqc_flags: rnaseqc_flags
       # gatk_flags: gatk_flags
     out:
@@ -309,7 +300,7 @@ steps:
   #   out: [out_hash_string]
 
 $namespaces:
-  s: https://schema.org/
+  s: http://schema.org/
 
 $schemas:
 - http://dublincore.org/2012/06/14/dcterms.rdf
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker.cwl
index e08d4da..718247b 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/checker-workflows/rnaseq_pipeline_fastq_checker.cwl
@@ -19,20 +19,16 @@ inputs:
     type: File[]
   prefix_str:
     type: string
-  threads:
-    type: int
-  memory:
-    type: int
   rsem_ref_dir:
     type: Directory
   max_frag_len:
     type: int
   estimate_rspd:
-    type: string
+    type: boolean
   is_stranded:
-    type: string
+    type: boolean
   paired_end:
-    type: string
+    type: boolean
   genes_gtf:
     type: File
   genome_fasta:
@@ -40,8 +36,6 @@ inputs:
     secondaryFiles:
       - .fai
       - ^.dict
-  java_path:
-    type: string
   rnaseqc_flags:
     type: string[]
   # gatk_flags:
@@ -71,8 +65,6 @@ inputs:
   #   type: string
   # hash_exon_counts:
   #   type: string
-  hash_count_metrics:
-    type: string
   # hash_count_outputs:
   #   type: string
   checker_star_output_bam:
@@ -144,8 +136,6 @@ steps:
       star_index: star_index
       fastqs: fastqs
       prefix_str: prefix_str
-      threads: threads
-      memory: memory
       rsem_ref_dir: rsem_ref_dir
       max_frag_len: max_frag_len
       estimate_rspd: estimate_rspd
@@ -153,7 +143,6 @@ steps:
       paired_end: paired_end
       genes_gtf: genes_gtf
       genome_fasta: genome_fasta
-      java_path: java_path
       rnaseqc_flags: rnaseqc_flags
       # gatk_flags: gatk_flags
     out:
@@ -297,7 +286,7 @@ steps:
   #   out: [out_hash_string]
 
 $namespaces:
-  s: https://schema.org/
+  s: http://schema.org/
 
 $schemas:
 - http://dublincore.org/2012/06/14/dcterms.rdf
diff --git a/topmed-workflows/TOPMed_RNAseq_pipeline/samtools-sort.cwl b/topmed-workflows/TOPMed_RNAseq_pipeline/samtools-sort.cwl
index 9980e15..887518d 100644
--- a/topmed-workflows/TOPMed_RNAseq_pipeline/samtools-sort.cwl
+++ b/topmed-workflows/TOPMed_RNAseq_pipeline/samtools-sort.cwl
@@ -27,10 +27,7 @@ inputs:
     type: int?
     inputBinding:
       prefix: -l
-    doc: |
-      Set compression level, from 0 (uncompressed) to 9 (best)
-
-    doc: Set number of sorting and compression threads [1]
+    doc: Set compression level, from 0 (uncompressed) to 9 (best)
   input:
     type: File
     inputBinding: