Midway Subread pipeline

These scripts process RNA-seq data using the Subread pipeline. The scripts are designed to be used with the Slurm scheduler on the RCC Midway2 cluster at the University of Chicago, but can be adapted to other computing infrastructure.

Setup

1. Install the R/Bioconductor packages biomaRt and Rsubread:

   source("https://bioconductor.org/biocLite.R")
   biocLite(c("biomaRt", "Rsubread"))

2. Clone this repository with git clone or download and unzip this zip file.

Pipeline

The scripts should be submitted in the following order. Always run them from the head node in the root of the project directory.

1. Download and index the genome with download-genome.R:

   sbatch --mem=12G --partition=broadwl download-genome.R

2. Download and format the exons with download-exons.R:

   sbatch --mem=2G --partition=broadwl download-exons.R

3. Submit mapping jobs with submit-subread.sh, which calls run-subread.R for each fastq file:

   bash submit-subread.sh

4. Submit counting jobs with submit-subread.sh, which calls run-featurecounts.R for each BAM file:

   bash submit-subread.sh

Directory structure

These scripts expect the following directory structure. Only fastq/ has to be created manually. The scripts will create bam/, counts/, and genome/.

.
├── bam
├── counts
├── fastq
│   ├── YG-172S-S8-8_S125_L006_R1_001.fastq.gz
│   └── YG-172S-S8-9_S126_L006_R1_001.fastq.gz
└─── genome

• fastq/ - Contains raw data in fastq.gz files.

• bam/ - Contains the mapped reads in BAM files.

• counts/ - Contains the read counts in text files.

• genome/ - Contains the index genome and the exons coordinates in SAF format.

License

The code is available under the CC0 license, i.e. you are free to do whatever you like with this code, but it comes with no guarantees. See the file LICENSE for full details.

More information

• RCC User Guide
• Gilad Lab Midway Guide