DSL1: Input file name collision when merging two already merged BAMs in `additional_library_merge` step.

[TCLamnidis]

Check Documentation

I have checked the following places for your error:

• nf-core website: troubleshooting
• nf-core/eager pipeline documentation
  - nf-core/eager FAQ/troubleshooting can be found here

Description of the bug

In niche cases where multiple UDG treatments exist for a sample, and multiple libraries have each of these treatments, a file name collision kills the pipeline at the additional_library_merge step.

Steps to reproduce

Steps to reproduce the behaviour:

1. Command line: nextflow run ... (any input that requires merging of two already merged library-level BAMs during additional_library_merge step.
2. See error:

Caused by:
  Process `additional_library_merge` input file name collision -- There are multiple input files for each of the following file names: MIS010_ss_libmerged.trimmed.bam, MIS010_ss_libmerged.trimmed.bam.bai

Expected behaviour

The BAMs initial libmerged should have unique names, to avoid such errors.

Log files

Have you provided the following extra information/files:

• The command used to run the pipeline
• The .nextflow.log file
• The exact error:

System

• Hardware: HPC
• Executor: sge
• OS: Ubuntu
• Version 20.04.6 LTS

Nextflow Installation

• Version: 21.10.6 build 5660

Container engine

• Engine: Singularity
• version: singularity version 3.7.1
• Image tag: 2.4.5

Additional context

[jfy133]

In niche cases where multiple UDG treatments exist for a sample, and multiple libraries have each of these treatments, a file name collision kills the pipeline at the additional_library_merge step.

That's what confuses me... shouldn't they have been merged at the first post-dedup merging step? 🤔

[TCLamnidis]

They are, that's the problem. as they end up with the same name. Could be an issue with the naming of the initial library merge step, OR the trimming step.

[TCLamnidis]

To give a better overview. Say we have a sample with 4 libraries with the following attributes:

Sample	Library	UDG_Treatment	Strandedness	Lane
ABC001	A0101	half	double	1
ABC001	A0102	half	double	1
ABC001	B0101	none	double	1
ABC001	B0102	none	double	1

The BAMs of the first two libraries will be merged at the initial lib_merge, and be named ABC001_udghalf_libmerged.bam.
Equally, the BAMs of the last two libraries will be merged at the initial lib_merge, and be named ABC001_udgnone_libmerged.bam.
However, once they undergo bam trimming, the outputs lose their UDG attribute, and both become ABC001_libmerged.bam
Once the two come together for the additional_library_merge step, the two input files share a name and the file collision pops up.