DSL2: Optional process execution based on file presence

[MB]

Hello all,

I want to do something which I thought would be easy to implement but for the life of me can't get to work... Some advise would be very welcome.

I want to make a workflow for Bowtie2 with an optional indexing process. The reference file will be provided as a parameter and within the workflow, I want to check if the index files (*.bt2) are present. If not, perform the building process first, if they are, continue to alignment.

At the moment, my solution fails on checking for a file. Is there a callable method for a file channel that returns true/false based on file presence? Already tried .exists() but get an error that this isn't a valid method.

workflow BOWTIE2 {

take:

reference // Channel: [reference]

reads // Channel: [samplename, reads]

main:

index = file("${reference.getParent() }/${reference.baseName }*.bt2")

(index_ch, no_index_ch) = ( index.isEmpty()

? [ Channel.empty(), tuple(reference, reference.getParent()) ]

: [ reference.combine(Channel.fromPath("${reference.getParent() }/${reference.baseName }*.bt2")) , Channel.empty() ] )

build(no_index_ch)

align(reads.combine(build.out.mix(index_ch)))

}

[Paolo Di Tommaso]

Think can be simplified a lot, with dsl2 you can just use an `if` statement. 

```

if( index.isEmpty() ) {

align( .. )

}

else {

build( .. )

align( .. )

}

```

Take it as pseudocode, I've omitted the combination of the corresponding channels.

p

--
You received this message because you are subscribed to the Google Groups "Nextflow" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nextflow+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nextflow/42a3f863-8925-4ed1-9163-b6c8d1a7d545n%40googlegroups.com.

[Mark Bessem]

Hey Paolo,

Thanks, that already makes my code a lot cleaner. I'm still stuck with checking for files at the original location of the reference file though. Is there a nextflow method for getting the parent of the original file in a channel? Otherwise I'd have to do it in a map function, which complicates things quite a bit since I'd have two possible output channels defined within a map function. .getParent() in the workflow doesn't seem to work for this purpose.

take:

   reference // Channel: [reference.fasta]

   reads // Channel: [samplename, reads]

main:

   index = file("${ reference.getParent() }/${ reference.baseName }*bt2")

   if( index.isEmpty() ) {

build( reference )

align( reads.combine(build.out) )

}

else {

align( reads.combine( tuple (reference, index ) )

}

Op vr 7 jan. 2022 om 03:16 schreef Paolo Di Tommaso <paolo.d...@gmail.com>:

To view this discussion on the web visit https://groups.google.com/d/msgid/nextflow/CADgKzdwVwos9Gh1KssB%3Dy-Jj6WVpf-KRsZFh%2B%2Bxsv%2BOPeziBbw%40mail.gmail.com.

[drhp...@gmail.com]

Hi Mark,

In terms of workflow design it may be an issue assuming that the reference and the index are in the same path as above. Another solution is to parameterise "params.bowtie2_index" and then to have some pre-processing to generate the index followed by alignment. An example of this is below:
https://github.com/nf-core/viralrecon/blob/2ebae61442598302c64916bd5127cf23c8ab5611/subworkflows/local/prepare_genome_illumina.nf#L115-L129

The code above doesn't explicitly check that the ".bt2" suffix files exist but the pipeline would fail at the alignment step if this isn't provided correctly.

Cheers,

Harshil