RNA-seq Nextflow pipeline and associated files

Setting up a new run

1. Create a run directory, change current directory to it, and retrieve files from GitHub to it using the command:

   git clone https://github.com/compbiomed/RNA_Seq ./

2. Generate a tab-delimited Nextflow input file following the format described below under params.infile.

3. Edit the RNA-seq_template.config file:

   • Set params.infile to the full path to the tab-delimited file describing the FASTQ input files.
   • Set params.output_dir to the full path to the Nextflow run directory.
   • Set params.prefix to a meaningful name for the project. This string will be used as a prefix to label many output files.
   • Set the fields of params.genome:
     • species: The scientific name of the species being used (e.g., "Homo sapiens", "Mus musculus")
     • ucsc: The UCSC build corresponding to the FASTA reference that will be used (e.g., "hg38", "mm10"), as generated, for example, by make_ucsc_references.qsub If an Ensembl genome reference is to be used instead, either remove this field or leave it set to "".
     • assembly: The corresponding genome assembly (e.g., "GRCh38", "GRCm38")
     • set: The subset of sequences that will be used: "base" (autosomes, sex chromosomes, and mitochondrial chromosome), "base_random" (base sequences plus random/unplaced contigs), or "base_random_althap" (base and random sequences plus alternative haplotype sequences); see make_ucsc_references.qsub for more details
     • ensembl: The Ensembl build number that will be used (e.g., 100)
   • Change params.read_length, params.paired_end, and params.stranded if needed (rare).

4. Rename the RNA-seq_template.config file to something more meaningful (e.g., [params.prefix].config)

5. Start the Nextflow run using the qsub file as follows:

   qsub -P [SGE project name] submit_RNA_Seq.qsub [config filename]

This will run the Nextflow script RNA_Seq.nf using the input file and config files from steps 2-4.

Description of files

RNA_Seq.nf

This Nextflow pipeline contains the following processes:

Setup

• generateGTF:
  • Retrieve a GTF file from Ensembl for the species, genome assembly, and Ensembl build number specified in the .config file
  • Use the UCSC chromAlias table for the corresponding UCSC genome build to convert the sequence names (i.e., chromosomes) from Ensembl to UCSC nomenclature
  • Limit the result to the subset of sequences specified in params.genome.set (e.g., "base_random")
• generateBED: Use the UCSC command-line utilities gtfToGenePred and genePredToBed to convert the GTF file to a BED file (for use with RSeQC below)
• runRSEMprepareReference: Prepare a set of RSEM reference files using the FASTA reference specified in the .config file and the Ensembl

FASTQ-level QC

• runFastQC: Use FastQC to perform QC on each pair of FASTQ files
• runMultiQCFastq: Compile output from runFastQC into TSV tables and HTML report using MultiQC

Alignments

• runSTAR1pass: Perform a first-pass alignment to a specified genome using STAR
• runSTARgenomeGenerate: Create a new genome reference from splice junctions inferred from first-pass STAR alignment
• runSTAR2pass: Perform a second-pass alignment to the genome reference produced by runSTARgenomeGenerate

BAM QC, performed using RSeQC

• runRSeQCbamStat
• runRSeQCclippingProfile
• runRSeQCdeletionProfile
• runRSeQCgeneBodyCoverage Note: this is a very slow step
• runRSeQCinferExperiment
• runRSeQCinnerDistance
• runRSeQCinsertionProfile
• runRSeQCjunctionAnnotation
• runRSeQCjunctionSaturation
• runRSeQCreadDistribution
• runRSeQCreadDuplication
• runRSeQCreadGC
• runRSeQCreadNVC
• runRSeQCreadQuality
• runRSeQCtin Note: this is a very slow step

Other BAM-level QC

• runMultiQCSample: Compile output from RSeQC and STAR using MultiQC

Post-alignment processing

• runRSEMcalculateExpression: Use RSEM to estimate gene- and transcript (isoform)-level expression
• createSE: Create SummarizedExperiment R objects, one for gene-level expression and one for transcript-level expression; each object contains several estimates of expression from RSEM, as well as QC parameters and feature-level annotation

RNA_Seq_template.config

This configuration file is intended to be used only with this Nextflow script. It makes a number of assumptions about underlying directory structures and filenames. The parameters that are typically changed are:

params.infile

Full path to a TSV file containing the following columns (those in bold are not optional and cannot be left blank):

• INDIVIDUAL_ID: An ID for an individual from which one or more samples was obtained (if only one sample was sequenced from each individual, this can be the same as SAMPLE_ID, or left blank)
• SAMPLE_ID: An ID for each sample
• LIBRARY_ID: An ID for each library prepared from a sample (if only one library was sequenced from each sample, this can be the same as SAMPLE_ID, or left blank)
• RG_ID: Read Group ID: the flowcell ID, optionally followed by a lane-specific suffix (for instruments with independent lanes), followed by a sample-specific identifier (e.g., SAMPLE_ID)
• PLATFORM_UNIT: RG ID, optionally followed by a suffix specific to a library (if more than one library was sequenced per sample)
• PLATFORM: Sequencing platform, e.g., "illumina" for Illumina instruments
• PLATFORM_MODEL: Instrument model, e.g., "NextSeq", "HiSeq", etc.)
• RUN_DATE: Optional run date
• CENTER: Optional name for center at which sequencing was performed
• R1: Full path to FASTQ file containing first paired-end read
• R2: Full path to FASTQ file containing second paired-end read

Note: some of these fields are discussed in more detail within the GATK read groups documentation.

params.output_dir

Full path where the Nextflow output should be written

params.prefix

Prefix for Nextflow output files

params.paired_end

Indicates whether paired-end sequencing was used (true or false)

params.genome

Parameters specific to the genome and annotation build used; these can be uncommented and edited as needed

submit_RNA_Seq.qsub

This script kicks off the Nextflow process on the SGE using the .config file specified in its sole argument.