https://nf-co.re logo
Join Slack
Powered by
And this one as well: <github.com/nf-core/sarek/pu...
# sarek_dev
f
m
I would replace the ext.when via .filter{}
f
feel free to improve
😆 1
it took me quite a while to set the output things right so that they produce exacttly the files you would think they do
r
I think we need to be careful with the parabricks haplotypecaller chaining here. if I understand the parabricks docs correctly running pb haplotypecaller already recalibrates on the fly
f
ok so that means 
--parabricks_applybqsr
should not be combined with pb hc? I think we can just catch that in the beginning of the workflow.
r
can you double check the docs. I just have a faint memory of when I reviewed that PR
👍 1
f
You can provide an optional BQSR report to fix the BAM, similar to ApplyBQSR. In this case, the updated base qualities will be used.
That means if we hand the bqsr report from fq2bam to haplotypecaller it does the recalibration itself. So we have two options: 1. hand over the bqsr report and make sure that the tools are mutually exclusive 2. just not hand the bqsr report to haplotypecaller (then it will not do recalibration)
Any preference? Because at the moment basically 2 is implemented (the sbwf for calling with haplotypecaller takes no bqsr table as input)
I am just wondering if parabricks applybqsr needs to be its own subworklfow (that is the case for spark and normal gatk) but seems not so logical to me
r
sorry haven't had time yet. will try to get to it later. otherwise I have a block for all the parabricks PRs tmrw morning
f
no worries i am not in a rush to merge
@Gary Burnett from the parabricks/applybqsr module it seems like the tool can take CRAM (at least there are tests for it), when I look at the official docs there only bam is mentioned. What is your take?
g
Hey @Famke Bäuerle, thank you for asking. The docs do not explicitly say it (I just filed a bug) but 
applybqsr
does support BAM and CRAM input.
woohoo 1
👌 1
f
Ok thats nice! What about the output? Can I tweak it to output crams?
r
this will make your logic so much simpler success
sarek deal with it 1
g
Yes it supports CRAM output as well!
Anything to make the logic simpler and bring it to parity with the current workflows!
f
Do I just need to set the filename or is there a flag for cram output?
g
Setting the output name should be enough
r
what cram encoding is used?
g
So it's a little disjointed right now. We support CRAM 3.1 for input and 3.0 for output.
😂 2
🤔 1
r
@Famke Bäuerle I think that is fine for the downstream tools
👍 1
f
ok let me adjust the PR, I think this should be documented somewhere because how would I know from reading the parabricks docs that its possible 😞
I think this PR should then allow to handle everything: github.com/nf-core/modules/pull/12765
👀 1
(still updating the snaps)
And if we have 
save_output_as_bam
we just use bam for all downstream? so no conversion?
r
should be: github.com/nf-core/sarek/pull/2154/changes we did a lot of back and forthy conversions which then didn't actually save much space
👍 1
f
ok PR would be ready for review I think I fixed most of the comments
6 Views
Open in Slack
PreviousNext
Powered by