Mitochondria workflow about raredisease

[jfy133]

I don't think we need a specific Mito dataset that can't be swapped out for another circular genome (e.g. bacteria).

That said all of the pipeline tests dataset is mammoth MiTo genomes.... But only because it was small.

Regardless I don't think we would anything special :)

Happy to give input on any discussion though

[maxulysse]

Yeah, @GCJMackenzie has wonderfully made many modules + test data for it, so plan would be to add it there soonish
cc @FriederikeHanssen

[FriederikeHanssen]

module for mutect2 is already in nf-core/modules https://github.com/nf-core/modules/tree/master/modules/gatk4/mutect2 It is much more comprehensive than my tiny attempt in sarek, so we will definitely use that one, just haven't gotten around to do this, but hopefully during the next couple of days there is time

[GCJMackenzie]

Might be best if I change -L chrM to a string input, that way its not hardcoded to use the 'chrM' label. This might also save some issues down the line if the pipeline tests are mammoth miTos and have a different labelling system.
The --mitochondria-mode seems to just change a few other settings to make the detection more sensitive, so if I am reading the gatk documentation correctly we might be able to just use a bacterial dataset (fingers crossed) :)
Will give it a go tomorrow.

[FriederikeHanssen]

sounds great. and some bacterial should already be in the testdata repo 🥳

[Gwennid]

The module for vep is called "ensemblvep".

[jemten]

Nice! Looking forward to this one

[Gwennid]

There is a local module for Mutect2 in Sarek (dev branch) so perhaps we can use it somehow.

[jemten]

Yeah, let's reuse as much code as possible.

@maxulysse, any plans of taking the mutect2 module to nf-core?

[Gwennid]

Thanks @FriederikeHanssen ! I had overlooked it...

[FriederikeHanssen]

no worries, but that is the beauty of the modules 😍

[GCJMackenzie]

Hi all, the module on nf-core currently does not have a specific mitochondrial mode, but I don't think it would be too tricky to add, if you needed it in a rush a workaround currently would be to run the generate_pon mode with the additional arguments -L chrM and --mitochondria-mode.

[GCJMackenzie]

@FriederikeHanssen @Gwennid I have started adding in a mitochondrial mode for mutect2. I don't suppose there is any test data for mitochondria yet is there?

[Gwennid]

I don't know of any.
I'd be happy to help, but I cannot promise much before mid-November unfortunately.

[FriederikeHanssen]

nope, not as far as i am aware and the keys don't indicate as such yet. Maybe the eager crowd knows more? @jfy133 anything like that on your side yet? Or would it be useful for you in the future? In that case it would probably be good to collect requirements from the start

[GCJMackenzie]

I think the changes to the module code itself are good, could upload a PR tomorrow with them. Unfortunately the tests fail since our current test data is for chr22 and not the mitochondria. (was hoping maybe mutect2 would give us a blank vcf but no luck, mutect2 doesn't like when -L specifies a chromosome not present in the file.)

[GCJMackenzie]

@jfy133 I think you are right :) I'd be happy to use a dataset that isn't human. The error being thrown at me is that there isn't a chrM labelled in my test files.

[FriederikeHanssen]

can we change the label in the bacteria dataset or will that cause some other unintended problem 🤔

[Gwennid]

@sysbiocoder , @Lucpen and @Gwennid have discussed how to structure the mitochondria subworkflow. At the moment we are thinking of separating it in three: one part for each of the two parallel "tracks" (see here) and a main part with the three initial steps, that prepares the inputs for the two parallel tracks, calls them, and then joins the outputs and runs the end (annotation etc).

[ramprasadn]

Closing this issue since this feature is already implemented in the pipeline.