for some mad reason the hisat2 module was directly...
# request-review
j
for some mad reason the hisat2 module was directly referencing
params
. Removed madness, made consistent with STAR: https://github.com/nf-core/modules/pull/10246
actually re-thinking...
m
I WANTED TO DO THE SAME
Thank for doing it before me
j
There are complications though, I'll get back to you 😉
m
so for me both seq_platform and seq_center should be coming from the meta map
as we do in the read group in sarek for example
j
agreed! Also means I need to fix star the same way
m
I shared some thoughts in the comments
but ALL for it
THANKS YOU SO MUCH
👍 1
If I didn't had the #C06PHJM0Y3D release today, I would have done it today
s
so for me both seq_platform and seq_center should be coming from the meta map
Yes, that would be sensible. And ideally they would be set via the module, rather using a config...
p
And ideally they would be set via the module, rather using a config...
ohh,, this isnt the recommended way?
Copy code
withName: STAR_ALIGN {
        ext.args = {
            return "........ --outSAMattrRGline 'ID:${meta.id}' 'SM:${meta.id}' 'LB:${meta.read_group_library}' 'PL:${meta.read_group_platform}' 'PU:${meta.read_group_platform_unit}'"
        }
    }
s
@Peter Kneale, that is indeed the current recommended way
✅ 1
j
I made another PR for Bowtie2 for consistency with this, basically adding some default read groups behaviour: https://github.com/nf-core/modules/pull/10253 .
🚀 1
✅ 1
Thanks @Maxime Garcia! This is just the workflow-level install of the same: https://github.com/nf-core/rnaseq/pull/1713/changes I wanted to allow sample-specific values for the tags to address the linked issue, which is what prompted all this.