output:
    tuple val(meta), path("${prefix}.bam"), emit: bam, optional: true
    tuple val(meta), path("${prefix}*bai"), emit: bai, optional: true
    tuple val(meta), path("${prefix}.cram"), emit: cram, optional: true
    tuple val("${task.process}"), val('gatk4'), eval("gatk --version | sed -n '/GATK.*v/s/.*v//p'"), topic: versions, emit: versions_gatk4

also this is the output of the module - why has it a bai but no crai then?

if [[ ${suffix} == cram ]]; then
        touch ${prefix}.cram.bai

shakefist

I don't think it'll solve the issue here, I just think we should put sam/bam/cram into one channel (and the index)

but not 100% sure

and solved if you just make the module emit the two things together

I really don't think it is a good idea to split them out

like the bai/csi separation

which means that sarek doesn't work for large genomes

Did you solve this? Because I tried tracing through the code, I’m still not entirely clear which

.join

is the culprit but this one seems like a possible culprit? I don’t know Sarek that well, but you don’t use the outputs of

FASTQ_PREPROCESS_GATK

if someone passes

save_to_bam

? (Well you mix the channel but AFAICT it would be empty)