BAM > featurecount > DEseq2 Nextflow/Snakemake pipeline

Question

[MJ500]

Hey guys,

typically I am using full DE Nextflow pipeline that as an input takes fastq files.

However now I want to compare my data with other people experiments from NCBI SRA. Unfortunately I have data in BAM format. I know that I could convert it back to fastq however it would be super inefficient for over 400 files.

Does anybody knows any Nextflow/Snakemake pipelines that accepts BAMs for DE DEseq2 analysis?

Thank you for any tips!

Answer 1

[dariober]

Does anybody knows any Nextflow/Snakemake pipelines that accepts BAMs for DE DEseq2 analysis?

For snakemake, but I guess also for nextflow, if you put the bam files where snakemake expects them, then snakemake will start the pipeline from there. It doesn't matter if the fastq files don't exist provided of course that they are not needed for anything other than creating the bam files.

For example, if the pipeline takes in input fastq/{name}.fastq and gives in output bam/{name}.bam, put your bam files in the bam/ directory and snakemake will pick up from there. This assumes your bam files are named {name}.bam. If they are not, rename or symlink them. ({name} is a wildcard that snakemake will replace with the actual sample names as appropriate).

[Devon Ryan]

N.B., this will depend a bit on how the workflow manager determines the list of output sample names. Otherwise you can touch fastq/{name}.fastq so it can determine the sample names. I don't know if nextflow has that issue, since it has a push rather than pull model.

[dariober]

Indeed - and you may want to touch the fastq files to set the timestamps to be older than the bam files.

[MJ500]

@Devon Ryan and @dariober

Thank you very much. I was thinking about doing this this way however the workflow I am using has so many error checks that basically I would need to rewrite a full workflow :/

But thanks in the worst case scenario I will do it this way.