RNA-Seq Workflow Tutorial

1. Make project directory

# the project directory contains specific GROUP name and current DATE
GROUP=XXX
DATE=`date +"%Y%m%d"`
mkdir ~/Project/${GROUP}_${DATE}

2. Clone the repository

cd ~/Project/${GROUP}_${DATE}
git clone https://github.com/bioxfu/RNA-Seq
cd RNA-Seq

3. Copy/Download the raw data

DATAPATH=/the/path/of/the/raw/data/on/HPC

# if you are working on the HPC, copy the raw data
./script/copy_rawdata.sh $DATAPATH

# if you are working on the local machine, download the raw data
./script/copy_rawdata.sh $DATAPATH --download

4. Rename the raw data

# dry run to check if mv command is correct
./script/rename_rawdata.sh --dry-run

# then do it 
./script/rename_rawdata.sh

5. Create config.yaml and Snakefile based on the examples

cp example/example.HPC.plant.config.yaml config.yaml
cp example/example.plant.Snakefile Snakefile
cp example/example_design_table.tsv design_table.tsv

# edit config.yaml and design_table.tsv

6. Initiate the project

source init.sh

7. Dry run the workflow to check any mistakes

./dry_run.sh

8. Run the workflow

# if you are working on the HPC
./run_HPC.sh

# if you are working on the local machine
./run.sh

# check the workflow progress in nohup.out file
tail nohup.log 

# check the jobs on HPC
qstat

# to delete all the jobs at once
qselect -u <username> | xargs qdel

# if you get the error: Directory cannot be locked.
snakemake --unlock 

9. Remove the temporary files

./clean.sh

10. Remove unwanted variation from RNA-Seq data

## remove batch effect using RUV
/cluster/home/xfu/R/3.5.1/bin/Rscript script/RUVSeq.R