From 3fe71fce12a370b6eeb27124cffc4efe471373be Mon Sep 17 00:00:00 2001
From: Dominik Brilhaus <dominik.brilhaus@hhu.de>
Date: Tue, 9 Aug 2022 15:57:57 +0200
Subject: [PATCH] cwl kallisto_index
---
runs/kallisto_index/README.md | 21 ++++++++++
runs/kallisto_index/kallisto_index.yml | 7 ++++
workflows/_dependencies/install_kallisto.md | 21 ++++++++++
workflows/_dependencies/install_r_packages.R | 43 ++++++++++++++++++++
workflows/kallisto_index.cwl | 33 +++++++++++++++
workflows/kallisto_index.sh | 31 ++++++++++++++
6 files changed, 156 insertions(+)
create mode 100644 runs/kallisto_index/README.md
create mode 100644 runs/kallisto_index/kallisto_index.yml
create mode 100644 workflows/_dependencies/install_kallisto.md
create mode 100644 workflows/_dependencies/install_r_packages.R
create mode 100644 workflows/kallisto_index.cwl
create mode 100755 workflows/kallisto_index.sh
diff --git a/runs/kallisto_index/README.md b/runs/kallisto_index/README.md
new file mode 100644
index 0000000..6143bf5
--- /dev/null
+++ b/runs/kallisto_index/README.md
@@ -0,0 +1,21 @@
+
+# cwltool README
+
+## First cd into the (runs) folder, with the .yml file
+
+cd /Users/dominikbrilhaus/03_DataPLANT_gitlab/samplearc_rnaseq/runs/kallisto_index
+
+## Let it flow
+
+```bash
+### store arc root (two levels up from here) as variable
+arc_root=$(echo ${PWD%/*/*})
+
+### replace arc root line in yml (specific to the machine from where this is run)
+### not sure, if this works on linux...
+sed -i '' "s|^arc_root:.*|arc_root: $arc_root|g" kallisto_index.yml
+
+### run with cwltool
+cwltool ../../workflows/kallisto_index.cwl kallisto_index.yml
+
+```
diff --git a/runs/kallisto_index/kallisto_index.yml b/runs/kallisto_index/kallisto_index.yml
new file mode 100644
index 0000000..1190016
--- /dev/null
+++ b/runs/kallisto_index/kallisto_index.yml
@@ -0,0 +1,7 @@
+cores: 1
+sh_script:
+ class: File
+ path: ../../workflows/kallisto_index.sh
+in_genome_ref: studies/TalinumGenomeDraft/resources/Talinum.gm.CDS.nt.fa
+out_folder: runs/kallisto_index
+arc_root: /Users/dominikbrilhaus/03_DataPLANT_gitlab/samplearc_rnaseq
diff --git a/workflows/_dependencies/install_kallisto.md b/workflows/_dependencies/install_kallisto.md
new file mode 100644
index 0000000..af072ad
--- /dev/null
+++ b/workflows/_dependencies/install_kallisto.md
@@ -0,0 +1,21 @@
+
+# Kallisto
+
+## Manual
+
+http://pachterlab.github.io/kallisto/manual.html
+
+## MacOS Installation
+
+```bash
+ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
+brew install kallisto
+```
+
+```bash
+git clone https://github.com/pachterlab/kallisto.git
+cd kallisto
+mkdir build
+cd build
+cmake .. -DUSE_HDF5=ON
+```
diff --git a/workflows/_dependencies/install_r_packages.R b/workflows/_dependencies/install_r_packages.R
new file mode 100644
index 0000000..ce4e643
--- /dev/null
+++ b/workflows/_dependencies/install_r_packages.R
@@ -0,0 +1,43 @@
+#!/usr/bin/env Rscript
+
+########################
+# Installation of R dependencies
+########################
+
+### sleuth installation
+
+if (!requireNamespace("BiocManager", quietly = TRUE)){install.packages("BiocManager")}
+if(!"BiocManager" %in% row.names(installed.packages())){BiocManager::install()}
+if(!"devtools" %in% row.names(installed.packages())){BiocManager::install("devtools")}
+
+# BiocManager::install("pachterlab/sleuth")
+# There's currently (early 2022) an issue with sleuth installation.
+# Using this workaround from https://github.com/pachterlab/sleuth/issues/259#issuecomment-1001076030
+
+if(!"sleuth" %in% row.names(installed.packages())){remotes::install_github("pachterlab/sleuth#260")}
+
+library(sleuth)
+
+### other packages
+
+library(tidyverse)
+library(jsonlite)
+library(openxlsx)
+
+required_packages <-
+ c('tidyverse', ## data wrangling and plotting
+ 'jsonlite', ## read/write json files
+ 'openxlsx') ## read/write xlsx files
+
+for(package in required_packages)
+{
+ ## Check if package is installed. If not, install
+ if(!package %in% row.names(installed.packages())){install.packages(package,
+ repos ="https://cran.uni-muenster.de/")}
+ # ## Check if package is up to date. If not, update
+ # update.packages(package, repos = "https://cran.uni-muenster.de/")
+ ## Load package
+ library(package, character.only = T)
+ print(paste("installed R package", package))
+}
+
diff --git a/workflows/kallisto_index.cwl b/workflows/kallisto_index.cwl
new file mode 100644
index 0000000..a6389de
--- /dev/null
+++ b/workflows/kallisto_index.cwl
@@ -0,0 +1,33 @@
+#!/usr/bin/env cwl-runner
+
+cwlVersion: v1.2
+class: CommandLineTool
+
+inputs:
+- id: sh_script
+ type: File
+ inputBinding:
+ position: 0
+- id: arc_root
+ type: string
+ inputBinding:
+ position: 1
+- id: in_genome_ref
+ type: string
+ inputBinding:
+ position: 2
+- id: out_folder
+ type: string
+ inputBinding:
+ position: 3
+
+outputs:
+- id: outdir
+ type:
+ type: array
+ items: Directory
+ outputBinding:
+ glob: $(runtime.outdir)/$(inputs.out_folder)
+
+baseCommand:
+- bash
\ No newline at end of file
diff --git a/workflows/kallisto_index.sh b/workflows/kallisto_index.sh
new file mode 100755
index 0000000..19e4ba7
--- /dev/null
+++ b/workflows/kallisto_index.sh
@@ -0,0 +1,31 @@
+#!/usr/bin/env bash
+
+### Build kallisto index
+
+################################################
+#### Read arguments from CLI
+################################################
+
+arc_root=$1
+in_genome_ref=$2
+out_folder=$3
+
+################################################
+#### Print version and citation to test kallisto
+################################################
+
+kallisto version
+kallisto cite
+
+################################################
+#### If it does not exist, create out dir
+################################################
+
+mkdir -p "$arc_root/$out_folder/"
+
+
+################################################
+#### Build kallisto index
+################################################
+
+kallisto index -i "$arc_root/$out_folder/kallisto_index" "$arc_root/$in_genome_ref"
--
GitLab