RNA-seq analysis

A shell pipeline for automatically analyzing all RNA-seq data in a folder based on STAR-RSEM and GATK

Introduction

• run_RNAseq_STAR.RSEM.sh includes QC, trimming, contamination screening, mapping, gene/isoform expression quantification, and summary steps.

• callSNP_RNAseq.sh is for SNP calling using the RNA-seq analysis results for each sample.

Downstream analyses, such as different expression analysis and pathway enrichment analaysis, are case specific, so are not included in this pipeline.

Configuration

The paths of some genome files and tools need to be modified accordingly:

• Modify the paths of genome STAR index and ref, fa, gtf, refSeq, and vcf (if use SNP calling) files.
• Make sure fastqc, multiqc, trim_galore, fastq_screen, STAR, RSEM, samtools are installed and can be ran by just typing their names, otherwise modify the paths of these tools in this pipeline script.
• Install Picard tools and set an environment variable/alias named "PICARD" for the path of picard.jar script, otherwise manually change "$PICARD" in this pipeline script to the full path of picard.jar script. The same with GATK, if use the SNP calling script.
• Add summarize_RNAseq_STAR.RSEM.pl to environment PATH or put it in /usr/local/bin/, otherwise modify the path of this perl file in this pipeline script.

Usage

run_RNAseq_STAR.RSEM.sh -g <hg19, mm10 or rn6> <-p> -i <path_of_inputs> -o <path_of_outputs> -t <threads>

callSNP_RNAseq.sh -g <hg19> -i <input_folder> -t <threads>

Arguments

• -g: set hg19, mm10 or rn6 as reference genome, default is hg19.
• -p: set this for paired-end data, default is for single-end data
• -i: directory of input fastq or fastq.gz files.
• -o: directory for output files.
• -t: set average number of threads for each sample in parallel analysis, must be integer, default is 1.

Outputs

run_RNAseq_STAR.RSEM.sh generates:

• A qc folder including QC results.
• A process folder including a summary table, gene/isoform expression count/FPKM/TPM tables, and subfolders of analysis results for each sample.

callSNP_RNAseq.sh generates SNP calling vcf files in each sample folder.

Contact

Ting Wang ([email protected]).