Hi everyone! Hope you’re all doing great. I’m currently writing a module for CIRI2. The input/output channels are ready, but I’ve hit a roadblock with writing the

nf-test

. CIRI2 strictly requires a SAM file generated by BWA-MEM using the

-T 19

flag. Since the

nf-core/test-datasets

repo doesn't host raw, uncompressed SAM files for Homo sapiens, I'm weighing my options according to nf-core guidelines: 1. Chain modules in `setup`: If I run

BWA_MEM

in the setup block, I will also need to chain

BWA_INDEX

before it since pre-computed BWA index directories aren't available for the human test genome. This feels quite heavy for a single module unit test. 2. Local asset: Should I just include a tiny, heavily-truncated mock SAM file directly inside the module's

tests/assets/

directory? Has anyone encountered a similar tool dependency issue during testing? What is the recommended or preferred approach by maintainers in this scenario? Any guidance would be highly appreciated. Thanks in advance!

Option 1 is fine. It’s quite common to run another module in the

setup

phase. Here’s an example: https://github.com/nf-core/modules/blob/master/modules/nf-core/angsd/contamination/tests/main.nf.test

If you don't want the hassle with mapping, couldn't you use a samtools view in the setup stage and start from a bam file?

Yes @Alexander Nater, That's also another way to test module. I'll see which method works best with CIRI2.