CODEWITHSUNDEEP
snakemake
aetobatus-narinari
aetobatus-narinari

Reputation: 27

snakemake output files as list/table

I have written this rule in my snakefile

rule extractfeat:
    input:
        '/path/to/file/{genome}.gbk'
    output:
        '{genome}_{locus_tag}_{gene}_{substrate}.fasta'
    shell:
        '''
        extractfeat {input} {output} -value {wildcards.genome}_{wildcards.locus_tag} -type CDS -describe product,locus_tag
        '''

I want to get all the output files (by output files I mean file=row) that are written in a separate file that looks like this:

genome  locus_tag   gene    substrate
PalbDSM11370    02121   susC    pululan
PalbDSM11370    02122   susD    pululan
PalbDSM11370    01210   susC    arabinan
PalbDSM11370    01209   susD    arabinan
PalbDSM11370    02015   susC    bglukan
PalbDSM11370    02016   susD    bglukan
PpalDSM17968    00934   susC    pululan
PpalDSM17968    00933   susD    pululan
PpalDSM17968    02229   susC    arabinan
PpalDSM17968 02228 susD arabinan
PpalDSM17968 01622 susC bglukan
PpalDSM17968 01623 susD bglukan
PREVCOP 05864 susC pululan
PREVCOP 05865 susD pululan
PREVCOP 05852 susC arabinan
PREVCOP 05851 susD arabinan
PREVCOP 05099 susC bglukan
PREVCOP 05098 susD bglukan
PREVCOP 03646 susC ksiloglukan
PREVCOP 03645 susD ksiloglukan
Psp.AGR2160 00839 susC ksiloglukan

Thanks

Upvotes: 0

Views: 62

Answers (1)

Troy Comi
Troy Comi

Reputation: 2079

You can add this as another rule, with inputs dependent on all the outputs of your generating rule:

rule tabulate:
    input: <ALL THE FASTA FILES>
    output: 'table.txt'
    run:
        wcs = glob_wildcards('{genome}_{locus_tag}_{gene}_{substrate}.fasta')
        with open(output[0], 'w') as outfile:
            outfile.write('genome\tlocus_tag\tgene\tsubstrate\n')  # header
            for row in zip(*wcs):  # order will match order in wildcard string
                outfile.write('\t'.join(row) + '\n')

Or if you already have lists of those wildcards you can write that directly instead of using glob_wildcards.

Upvotes: 1

Related Questions