atpoint/rnaseq_preprocess repository overview

⁠rnaseq_proprocess

A Nextflow pipeline for preprocessing of RNA-seq data

⁠Workflow

• create a genome-decoyed index with salmon. This requires a reference transcriptome and genome, which can be provided as files from disk or by providing links to a server, e.g. an FTP address. Alternatively, the user can directly provide a premade index.

• trimming of the fastq files with cutadapt. This is not activated by default, use --trimming to turn it on. Default adapter sequence is the TruSeq adapter.

• quantification with salmon against the created or provided index.

• aggregation of the transcript abundance estimates to the gene level with tximport

All individual steps can optionally be turned off.

⁠Usage

Summary of important options:

• --fastq: path to fastq file (pairs), e.g. --fastq path/to/*.fastq.gz for single-end or --fastq path/to/*_{1,2}.fastq.gz for paired-end data. The data are expected to be gzip-compressed with suffix fastq.gz.

• --mode: either "single" or "paired" depending on the sequencing type.

• --idx: path to a premade salmon index. We expect salmon v1.0.0 and later.

• --ref_genome and --ref_txtome: paths to fasta files for genome and transcriptome (expected to end .fa.gz) or download links, e.g. FTP address to pull it from remote.

• --trimming: if set then perform adapter trimming with cutadapt

• --quant_libtype: the library type flag for salmon⁠, by default 'A' for automatic detection. Standard stranded libraries, e.g. NEBnext UltraII Directional Prep Kit would be 'ISR' and unstranded libraries are usually 'IU'.

All other options are intuitively named in the nextflow.config file. Some processes have an additional parameter to pass further arguments to the respective tool. If so the first parameter must be escaped to be parsed properly, e.g.:
--quant_additional '\--noLengthCorrection --numGibbsSamples 64'

For testing there are two profiles, -profile test_single and -profile test_paired for paired- and single-end data. Minimal example data⁠ are provided in ./test, which are also used by the GitHub Actions testing.

There are three profiles available to automatically take care of software dependencies which are -profile conda/docker/singularity.
Once Nextflow makes a stable release that supports mamba this profile will be added as well. The environment.yml contains the required software for conda.
A Docker image is available for the docker,singularity profiles at the Docker Hub⁠.

For submission via SLURM on a cluster one can use -profile slurm. The options --queue and --clusteroptions can be used to specify a queue and any other options the scheduler may use.

⁠Citations

• nf-core project from which much code inspiration was taken from⁠

• Ewels et al (2020) The nf-core framework for community-curated bioinformatics pipelines⁠

• The Nextflow Documentation⁠

• Seqera Training⁠

• https://github.com/nextflow-io/rnaseq-nf⁠

• Merkel, D (2014). Docker: lightweight linux containers for consistent development and deployment. Linux Journal⁠

• Kurtzer et al (2017) Singularity: Scientific containers for mobility of compute. PLoS ONE⁠

• Patro et al (2017) Salmon: fast and bias-aware quantification of transcript expression using dual-phase inference⁠

• Soneson et al (2015) Differential analyses for RNA-seq: transcript-level estimates improve gene-level inferences⁠