From 8bc5ebe548f1252ad58867ec0c61b7aa4a7212d8 Mon Sep 17 00:00:00 2001
From: Dominik Brilhaus <brilhaus@nfdi4plants.org>
Date: Fri, 2 Aug 2024 13:59:32 +0200
Subject: [PATCH] somewhat generic kallisto with multiple samples worked
---
.gitignore | 3 +++
runs/kallisto/README.md | 1 -
runs/kallisto/workflow.yml | 21 +++++++++++++++------
workflows/kallisto/kallisto-quant.cwl | 2 +-
workflows/kallisto/kallisto-workflow.cwl | 13 ++++++++-----
5 files changed, 27 insertions(+), 13 deletions(-)
diff --git a/.gitignore b/.gitignore
index 11819f2..bce78f5 100644
--- a/.gitignore
+++ b/.gitignore
@@ -386,3 +386,6 @@ Temporary Items
# R temporaries
.Rhistory
.RData
+
+# CWL caches
+**/cache*/
\ No newline at end of file
diff --git a/runs/kallisto/README.md b/runs/kallisto/README.md
index 65d21cd..2f01d07 100644
--- a/runs/kallisto/README.md
+++ b/runs/kallisto/README.md
@@ -6,4 +6,3 @@
cd runs/kallisto/
cwltool --cachedir cache2 ../../workflows/kallisto/kallisto-workflow.cwl workflow.yml > $(date +"%Y-%m-%d_%H-%M").log 2>&1 &
```
-
diff --git a/runs/kallisto/workflow.yml b/runs/kallisto/workflow.yml
index e1d32d5..db33ad6 100644
--- a/runs/kallisto/workflow.yml
+++ b/runs/kallisto/workflow.yml
@@ -1,4 +1,14 @@
-InputReadsMultipleSamples:
+## Genome file to build kallisto Index
+IndexInput:
+ - class: File
+ path: ../../studies/TalinumGenomeDraft/resources/Talinum.gm.CDS.nt.fa
+ format: edam:format_1929 # FASTA
+
+## Fastq files to be mapped
+# The `InputReadsMultipleSamples` looks more complicated than needed
+# It's an array of File arrays to generically allow multiple fastq files per sample
+
+InputReadsMultipleSamples:
-
- class: File
path: ../../assays/Talinum_RNASeq_minimal/dataset/DB_097_CAMMD_CAGATC_L001_R1_001.fastq.gz
@@ -7,14 +17,13 @@ InputReadsMultipleSamples:
- class: File
path: ../../assays/Talinum_RNASeq_minimal/dataset/DB_099_CAMMD_CTTGTA_L001_R1_001.fastq.gz
format: edam:format_1930 # FASTQ
-IndexInput:
- - class: File
- path: ../../studies/TalinumGenomeDraft/resources/Talinum.gm.CDS.nt.fa
- format: edam:format_1929 # FASTA
+
+numCharsFastqFileName: 6
+### Kallisto quant Parameters
isSingle: true
FragmentLength: 200
StandardDeviation: 20
-BootstrapSamples: 30
+BootstrapSamples: 1
$namespaces:
edam: https://edamontology.org/
\ No newline at end of file
diff --git a/workflows/kallisto/kallisto-quant.cwl b/workflows/kallisto/kallisto-quant.cwl
index 8267499..9d6df5f 100755
--- a/workflows/kallisto/kallisto-quant.cwl
+++ b/workflows/kallisto/kallisto-quant.cwl
@@ -132,6 +132,6 @@ outputs:
glob: $(runtime.outdir)/$(inputs.QuantOutfolder)
$namespaces:
- edam: http://edamontology.org/
+ edam: https://edamontology.org/
$schemas:
- https://edamontology.org/EDAM_1.18.owl
diff --git a/workflows/kallisto/kallisto-workflow.cwl b/workflows/kallisto/kallisto-workflow.cwl
index 37fb7df..c0e074e 100644
--- a/workflows/kallisto/kallisto-workflow.cwl
+++ b/workflows/kallisto/kallisto-workflow.cwl
@@ -7,13 +7,14 @@ requirements:
InlineJavascriptRequirement: {}
inputs:
+ IndexInput: File[]
InputReadsMultipleSamples:
type:
type: array
items:
type: array
items: File
- IndexInput: File[]
+ numCharsFastqFileName: int?
isSingle: boolean
FragmentLength: double?
StandardDeviation: double?
@@ -30,13 +31,15 @@ steps:
out: [index]
quant:
run: kallisto-quant.cwl
- scatter: InputReads
+ scatter:
+ - InputReads
+ - QuantOutfolder
+ scatterMethod: dotproduct
in:
InputReads: InputReadsMultipleSamples
QuantOutfolder:
- # source: InputReadsMultipleSamples
- # valueFrom: $(self[0].nameroot)
- valueFrom: testSample
+ source: InputReadsMultipleSamples
+ valueFrom: $(self[0].nameroot)
Index: index/index
isSingle: isSingle
FragmentLength: FragmentLength
--
GitLab