Repository navigation
Add SAVANA, padfoot and ReConPlot #192
Description
Description of feature
Hi,
I am currently working on adding SAVANA, padfoot and ReConPlot modules to the workflow as I found that the CN calling from ascat is not quite aligning to the SV reported as well as to the paired short reads.
Currently, I am testing on our cluster and will update with the progress if you think those modules could be beneficial.
Best,
Boyu
Activity
ljwharbers commentedon Sep 9, 2026
Hi Boyu!
That would be amazing. We currently have a PR open to implement SAVANA #189, so it would be great if you could have a look at the current implementation there and see if you would require any changes there. In terms of padfoot and ReConPlot I think these would be nice additions. What do you think? @robert-a-forsyth
Best,
Luuk
Tim-Yu commentedon Sep 10, 2026
Thank you. Will see how #189 cope with the other modules. I just found there is a nf-core savana. I was making a local version.
Correct the plugin documentation, and document the CSQ fields
Merge upstream/dev into vep_plugins
ljwharbers commentedon Sep 22, 2026
Hi Boyu,
Savana is now implemented. Padfoot and ReConPlot are not yet on our to-do but I can put it on there. you are of course more than welcome to help with the implementation and open a PR for it as well :D
Best,
Luuk
Tim-Yu commentedon Sep 22, 2026
Hi Luuk,
Just saw the PR. I will then modify my version for this Savana module then. Will PR later for the Padfoot and ReConPlot.
Best wishes,
Boyu
Tim-Yu commentedon Sep 23, 2026
Hi Luuk,
I took some time to rebase my Padfoot and ReConPlot work on top of the savana PR #189 and opened it as a draft PR #202 so you can look before I mark it ready.
Padfoot annotates somatic SVs + CNAs and ReConPlot draws the figures, each once per caller pair that produced output for a sample (Severus + Wakhan; ASCAT + Severus; SAVANA SV and CNV on its own), paired and tumour-only, on by default with
--skip_padfoot/--skip_reconplot.The only touch on the #189 code is three extra emits on the two savana subworkflows. As I was using my own implementation of savana, so that the Padfoot and ReConplot will need some file fixing. Separately, I will propose a few things for the savana stack that were in my local SAVANA module (branch
feat/sv-cna-annotation-visualisation) but are not in #189: deriving--chromosomesfrom the contigs list and a few unexposed CNA knobs(blacklist, bin size).The added modules are tested end to end on a chr22 paired ONT tumour/normal and the same tumour as tumour-only using singularity.
Currently I am running a full length real data testing to fine tuning the resource label for those.
Best,
Boyu
ljwharbers commentedon Sep 23, 2026
Awesome, thanks! Let me know when the full test runs through successfully and I'll have a look at the PR.
Tim-Yu commentedon Sep 23, 2026
Hi Luuk, details and the resource numbers are in a comment on #202. I just marked it ready for review.
ljwharbers commentedon Sep 24, 2026
Thanks! I've added some comments to the PR
Tim-Yu commentedon Sep 25, 2026
Thank you! I will fix on Monday if that is okey. Promised my wife no work today or weekends.....
ljwharbers commentedon Sep 25, 2026
That is more than okay haha, enjoy the day + weekend! 😂
Tim-Yu commentedon Sep 30, 2026
Thanks, fixed those now.
Meanwhile, I wonder if and how are you plan to include those to the final report and how your view on the arrangement of those new contents. I am making a version based on the current version.
Meanwhile, I wonder if you have a date for 1.2.0 release as I want to use as a anchor version for a lager scale running.
ljwharbers commentedon Sep 30, 2026
Thanks! I am not super familiar with both packages so would need to see how these plots look like and how to best fit them in. We wouldn't want to have too much in the final report because this will also ultimately go to clinicians.
Regarding 1.2.0 release, I think we want to have the current open PRs in for sure and maybe @robert-a-forsyth still wants to add shattersheek/other things? But in my opinion we can look into releasing it as 1.2.0 soon.