RNA-Seq tools

A collection of scripts to generate bash scripts on an SGE cluster for the following tools:

• STAR
• kallisto
• Tophat2
• Cufflinks/Cuffquant
• featureCount (subread)
• salmon
• SpliceTrap
• indexes (aligner index files)

Code

• scripts/ - python classes that wrap RNA-Seq tools/programs
• bin/ - scripts that use python classes to create qsub scripts
• templates/ - mako template for the qsub script

Dependencies

• Python 2
• mako template class
• Sun Grid Engine

Data organization

I assume the raw data (paired end FASTQs) has this layout:

runDirectory
|
|
└───sampleA
│   │
│   └───00-raw
│       │   sampleA_1.fastq.gz
│       │   sampleA_2.fastq.gz
│   
└───sampleB
│   │
│   └───00-raw
│       │   sampleB_1.fastq.gz
│       │   sampleB_2.fastq.gz
    

The data for each tool (STAR, kallisto, ...) will be output in its own subdirectory (eg 03-alignSTAR).

All rnaseq runs are assumed to be demultiplexed.

Config

appconfig.json - specify the path to the various software siteconfig.json - specify the base path for apps and indexes

Usage

To run a tool such as the STAR aligner, we create a JSON file in the run directory.

# config_STAR.json
{
"samples": [ "sampleA", "sampleB"],
"basedir": "/data/ranseq/run1/",
"reference": "ensembl_hg19",
"date": "2019-03-31"
}

To generate the scripts:

python make_STAR.py config_STAR.json

which will create files:

sampleA/sampleA_STAR_ensembl_hg19.sh
sampleB/sampleB_STAR_ensembl_hg19.sh

We can qsub all the files like this (from the run directory) find -type f -name *STAR*.sh -exec qsub {} \;