snakemake define folder as output

问题

I try to run prokka using snakemake and rule all. In the latter I define all output folders which will be produced by prokka to write the results. Prokka requires a folder to be supplied as an output rather than a file.

A simplified version of what I have is here:

PATIENTID_ls = range(2)
rule all:
input:
    expand("results_{subjectID}_outputfolder",subjectID=PATIENTID_ls), 

rule prokka:
    input:
        "contigs/subject_{subjectID}/contigs.fasta",
    output:
        "results/subject_{subjectID}_outputfolder",
    shell:
        "prokka --cpus 1 --proteins ../GCF_000009645.1_ASM964v1_genomic.gbff --outdir {output} --prefix contigs500_anno9ref {input} "

When running:

$snakemake -p
Building DAG of jobs...
MissingInputException in line 2 of Snakefile:
Missing input files for rule all:
results_1_outputfolder
results_0_outputfolder

It works however when specifying the output explicitly:

snakemake -p results/subject_1_outputfolder

I am sure that is noob mistake on my side, but after hours of playing around I could not solve the issue. Help is highly appreciated. Thank you

回答1:

Your example has some issues as your rule all output files do not match your rule prokka output file.

However, one way to implement what you want to do is to use params to specify the output directory and use that as argument to the flag --outdir {params.outdir}.

A similar example is shown below:

PATIENTID_ls = [1,2]
PREFIX = "contigs500_anno9ref"

rule all:
    input:
        expand("results_{subjectID}_outputfolder/{prefix}.gff",subjectID=PATIENTID_ls, prefix=PREFIX), 

rule prokka:
    input:
        "contigs/contigs.fasta",
    params:
        outdir= "results_{subjectID}_outputfolder",
        prefix= PREFIX,
    output:
        "results_{subjectID}_outputfolder/{prefix}.gff",
    shell:
        "echo '{params.prefix}' > {params.outdir}/{PREFIX}.gff"

You still should specify a file as an output in rule prokka and in rule all. Based on the example on the prokka repo the output file is essentially {outdir}/{prefix}.gff. You can specify that as the output to both rule all and rule prokka without ever directly using it while invoking the command.

Alternatively even though there does not seem to be a reason for it, you could use a mock file to signify completion of the rule.

An example would be:

PATIENTID_ls = [1,2]
rule all:
    input:
        expand("results_{subjectID}_outputfolder/mockfile.txt",subjectID=PATIENTID_ls), 

rule prokka:
    input:
        "contigs/contigs.fasta",
    params:
        outdir= "results_{subjectID}_outputfolder",
        prefix= "contigs500_anno9ref",
    output:
        "results_{subjectID}_outputfolder/mockfile.txt",
    shell:
        "echo '{params.prefix}' && touch {params.outdir}/mockfile.txt"

回答2:

As @JohnnyBD mentioned, your major problem appears to be with rule all output not matching rule prokka. If you still need to use directory as output instead of a file, you may want to use directory() as it better handles edge cases.

来源：https://stackoverflow.com/questions/53749519/snakemake-define-folder-as-output

标签

snakemake