Tutorial

This tutorial shows how convert the basic rnaseq-nf pipeline to Nextflow DSL-2.

Get started

Clone this repository in your computer, then change in the repo directory

https://github.com/nextflow-io/nfcamp-tutorial.git \
&& cd nfcamp-tutorial

Pull the required container:

docker pull nextflow/rnaseq-nf:latest

Change the standard profile in the nextflow.config as follow:

standard {
    process.container = 'nextflow/rnaseq-nf:latest'
    docker.enabled = true
    resume = true
}

Run the following script checking the differences each others.

Step 1 - Use of path and tuple input/output qualifier

nextflow run main1.nf 

Step 2 - Show use of yaml file

nextflow run main2.nf -params-file reads.yml

Step 3 - Show DLS-2 process and workflow

nextflow run main3.nf

Step 4 - Multiple use of the same channel

nextflow run main4.nf

Step 5 - Creation of a module files:

save the process to the file rnaseq-processes.nf

Step 6 - Inclusion of module file

nextflow run main6.nf

Step 7 - Selection process inclusion

include index from './rnaseq-processes' params(params)
include quant from './rnaseq-processes' params(params)
include fastqc from './rnaseq-processes' params(params)
include multiqc from './rnaseq-processes' params(params)

Step 8 - Create a sub-workflow module file named: rnaseq-analysis.nf

workflow rnaseq_analysis {
    get: 
        transcriptome
        read_pairs_ch

    main:
        index( transcriptome )
        
        quant( index.out, read_pairs_ch )
        
        fastqc( read_pairs_ch )
        
        multiqc( 
                quant.out.mix(fastqc.out).collect(),  
                params.multiqc )

}

Step 9 - Create two sub-workflows in the main script and use -entry to execute them

workflow rnaseqForTranscrip1 {
    rnaseq_analysis ( 
        params.transcriptome, 
        Channel .fromFilePairs( params.reads, checkExists: true )  )
}

workflow rnaseqForTranscrip2 {
    rnaseq_analysis ( 
        params.transcriptome, 
        Channel .fromFilePairs( params.reads, checkExists: true )  )
}



nextflow run main9.nf -entry rnaseqForTranscrip1
nextflow run main9.nf -entry rnaseqForTranscrip2

Step 10 - Use two different genome files

params.transcript1 = "$baseDir/data/ggal/transcriptome_1.fa"
params.transcript2 = "$baseDir/data/ggal/transcriptome_2.fa"

invoke both

workflow {
    rnaseqForTranscrip1()
    rnaseqForTranscrip2()
}

Step 11 - Use of the fork operator

workflow {
    reads = Channel .fromFilePairs( 'data/ggal/ggal_*_{1,2}.fq' ) 
    transcripts  = Channel.fromPath('data/ggal/transcriptome_*.fa')
    transcripts
        .combine( reads )
        .fork { tuple -> 
        trascript: tuple[0]
        reads: [ tuple[1], tuple[2] ]
        }
        .set { fork_out }
        
    rnaseq_analysis(fork_out)
}

Step 12 - Use of pipes

workflow {

    Channel .fromFilePairs( 'data/ggal/ggal_*_{1,2}.fq' ).set {reads} 
    Channel.fromPath('data/ggal/transcriptome_*.fa') \
        | combine( reads ) \
        | fork { tuple -> 
            trascript: tuple[0]
            reads: [ tuple[1], tuple[2] ]
        } \
        | rnaseq_analysis

}

Step 13 - Use of forkCriteria

workflow {
    separateTranscriptFromReads = forkCriteria({ tuple -> 
        trascript: tuple[0]
        reads: [ tuple[1], tuple[2] ]
        })

    Channel.fromFilePairs( 'data/ggal/ggal_*_{1,2}.fq' ).set {reads} 
    Channel.fromPath('data/ggal/transcriptome_*.fa') \
        | combine( reads ) \
        | fork(separateTranscriptFromReads) \
        | rnaseq_analysis
}

Step 14 - Use a custom function

def getInputForRnaseq( transcriptsPath, readsPath ) {

    def separateTranscriptFromReads = forkCriteria({ tuple -> 
        trascript: tuple[0]
        reads: [ tuple[1], tuple[2] ]
        })

    def reads = Channel.fromFilePairs(readsPath) 
    Channel.fromPath(transcriptsPath) \
        | combine( reads ) \
        | fork(separateTranscriptFromReads) 

}

workflow {
    getInputForRnaseq(params.transcripts, params.reads) | rnaseq_analysis
}

Step 15 - Use of workflow publish

• Remove publishDir from processes

• Add emit/out to rnaseq_analysis

• Add publish to the main workflow

    workflow {
        main:
        getInputForRnaseq(params.transcripts, params.reads) | rnaseq_analysis
        publish:
        rnaseq_analysis.out.fastqc to: 'results/fastqc_files'
        rnaseq_analysis.out.quant to: 'results/quant_files'
        rnaseq_analysis.out.multiqc to: 'results/multiqc_report'
    }