demultiplex icon indicating copy to clipboard operation
demultiplex copied to clipboard

Demux dies for empty files

Open FelixKrueger opened this issue 1 year ago • 2 comments

Description of the bug

On a number of occasions we have had the situation that the demux pipeline encounters the case where a sample in the sample sheet doesn't get any reads assigned to it, and the whole process dies. The 'fix' for this is to exclude the offensively empty sample from the samplesheet, and run again. This is a painful exercise, and for single-cell experiments with 384 barcodes per sample it can become unworkable.

Would it be possible to modify the workflow so that empty samples just appear as having a count of 0 reads in a summary table, and downstream processes that fail the entire pipeline are simply skipped?

Many thanks!

Command used and terminal output

No response

Relevant files

No response

System information

No response

FelixKrueger avatar Sep 26 '23 15:09 FelixKrueger

We have the same issue with @GiampaoloMart, the error is at line 125 of bcl_demultiplexing/main.nf:

    **assert line.startsWith('@')**
    line = line.substring(1)
    def fields = line.split(':')
    def rg = [:]

The assertion fails because the file is empty. This happens in the Fastp and MultiQC processes.

arnaudceol avatar Feb 15 '24 13:02 arnaudceol

Hi, same issue here, currently I am working in a fix, I still need to update the nf-core tests and pass them, hopefully soon it will succeed and be available in the dev branch (fingers crossed), you can follow at https://github.com/nf-core/modules/pull/4842

glichtenstein avatar Feb 18 '24 23:02 glichtenstein

Looks like #4842 is now merged.

k1sauce avatar May 14 '24 00:05 k1sauce

This is fixed by #186

k1sauce avatar May 15 '24 20:05 k1sauce