README.qmd

---
title: RNAseq.vsh
format: gfm
---

<!-- README.md is generated by running 'quarto render README.qmd' -->

```{r, echo = FALSE, message = FALSE, error = FALSE, warning = FALSE}
library(tidyverse)
```

A version of the [nf-core/rnaseq](https://github.com/nf-core/rnaseq) pipeline (version 3.14.0) in the [Viash framework](http://www.viash.io).

## Rationale

We stick to the original nf-core pipeline as much as possible. This also means that we create a subworkflow for the 5 main stages of the pipeline as depicted in the [README](https://github.com/nf-core/rnaseq).

## Getting started

As test data, we can use the small dataset nf-core provided with [their `test` profile](https://github.com/nf-core/test-datasets/blob/rnaseq3/samplesheet/v3.10/samplesheet_test.csv): <https://github.com/nf-core/test-datasets/tree/rnaseq3/testdata/GSE110004>.

A simple script has been provided to fetch those files from the github repository and store them under `testData/minimal_test` (the subdirectory is created to support `full_test` later as well): `bin/get_minimal_test_data.sh`. 

Additionally, a script has been provided to fetch some additional resources for unit testing the components. Thes will be stored under `testData/unit_test_resources`: `bin/get_unit test_data.sh`

To get started, we need to:

1.  [Install `nextflow`](https://www.nextflow.io/docs/latest/getstarted.html) system-wide

2.  Fetch the test data:

``` bash
bin/minimal_test.sh
bin/get_minimal_test_data.sh
```

## Running the pipeline

To actually run the pipeline, we first need to build the components and pipeline:

``` bash
viash ns build --setup cb --parallel
```

Now we can run the pipeline using the command:

``` bash
nextflow run target/nextflow/workflows/pre_processing/main.nf \
  -profile docker \
  --id test \
  --fastq_1 testData/minimal_test/input_fastq/SRR6357070_1.fastq.gz \
  --publish_dir testData/test_output/
```

Alternatively, we can run the pipeline with a sample sheet using the built-in `--param_list` functionality: (Read file paths must be specified relative to the sample sheet's path)

``` bash
cat > testData/minimal_test/input_fastq/sample_sheet.csv << HERE
id,fastq_1,fastq_2,strandedness
WT_REP1,SRR6357070_1.fastq.gz;SRR6357071_1.fastq.gz,SRR6357070_2.fastq.gz;SRR6357071_2.fastq.gz,reverse
WT_REP2,SRR6357072_1.fastq.gz,SRR6357072_2.fastq.gz,reverse
RAP1_UNINDUCED_REP1,SRR6357073_1.fastq.gz,,reverse
HERE

nextflow run target/nextflow/workflows/rnaseq/main.nf \
  --param_list testData/minimal_test/input_fastq/sample_sheet.csv \
  --publish_dir "test_results/full_pipeline_test" \
  --fasta testData/minimal_test/reference/genome.fasta \
  --gtf testData/minimal_test/reference/genes.gtf.gz \
  --transcript_fasta testData/minimal_test/reference/transcriptome.fasta \
  -profile docker
```

## Pipeline sub-workflows and components

The pipeline has 5 sub-workflows that can be run separately.

1. Prepare genome: This is a workflow for preparing all the reference data required for downstream analysis, i.e., uncompress provided reference data or generate the required index files (for STAR, Salmon, Kallisto, RSEM, BBSplit).

2. Pre-processing: This is a workflow for performing quality control on the input reads It performs FastQC, extracts UMIs, trims adapters, and removes ribosomal RNA reads. Adapters can be trimmed using either Trim galore! or fastp (work in progress).

3. Genome alignment and quantification: This is a workflow for performing genome alignment using STAR and transcript quantification using Salmon or RSEM (using RSEM's built-in support for STAR) (work in progress). Alignment sorting and indexing, as well as computation of statistics from the BAM files is performed using Samtools. UMI-based deduplication is also performed. 

4. Post-processing: This is a workflow for duplicate read marking (picard MarkDuplicates), transcript assembly and quantification (StringTie), and creation of bigWig coverage files.

5. Pseudo alignment and quantification: This is a workflow for performing pseudo alignment and transcript quantification using Salmon or Kallisto. 

6. Final QC: This is a workflow for performing extensive quality control (RSeQC, dupRadar, Qualimap, Preseq, DESeq2, featureCounts). It presents QC for raw reads, alignments, gene biotype, sample similarity, and strand specificity (MultiQC).

## Reusing components from biobox
At the moment, this pipeline makes use of the following components from [biobox](https://github.com/viash-hub/biobox):

* `gffread`
* `star/star_genome_generate`
* `star/star_align_reads`
* `salmon/salmon_index`
* `salmon/salmon_quant`
* `featurecounts`
* `samtools/samtools_sort`
* `samtools/samtools_index`
* `samtools/samtools_stats`
* `samtools/samtools_flagstat`
* `samtools/samtools_idxstats`
* `multiqc` (work in progress - updating `assets/multiqc_config.yaml`)
* `fastp` (work in progress)
* `rsem/rsem_prepare_reference` (work in progress)
* `rsem/rsem_calculate_expression` (work in progress)