Clinical-Genomics
[User Story] Use nf-core/createpanelref for PON generation
Need
As a developer I want to generate all BALSAMIC panel of normals using the nf-core createpanelrefs pipeline so I don't have to maintain PON generation workflows in BALSAMIC and we can harmonise how PONs are created throughout Clinical Genomics
Suggested approach
We should identify all PONs which we currently have workflows for in BALSAMIC and make modules for them using the NF-core createpanelrefs pipeline instead. Once that is in place we should create a workflow for how we create new PONs using this pipeline.
We should consider also new PONs being considered such as #1115, #1135 and #1368 and make sure we have a clear strategy for how we sync these two efforts up. Just to make sure we are not making more workflows in BALSAMIC if we already intend on replacing them
Considered alternatives
The alternative is to keep making workflows in BALSAMIC. The up-side to this is that we have a tighter control over how the PONs are generated.
The downside to this is that we keep expanding a pipeline which is in line to be replaced. The day that we do switch from BALSAMIC to another pipe, if this work is done we won't have to also create a new PON generation workflow.
Deviation
No response
System requirements assessed
- [ ] Yes, I have reviewed the system requirements
Requirements affected by this story
No response
Risk assessment needed
- [X] Needed
- [ ] Not needed
Risk assessment
No response
SOUPs
As this would outsource the work of PON generation to another pipeline, we should perhaps consider nf-core createpanelrefs as a SOUP and do a description and risk assessment of it.
Can be closed when
- [ ] The feasibility to switching our PON generation to nf-core/createpanelrefs have been though through
- [ ] All current needed PONs are possible to generate through nf-core/createpanelrefs
- [ ] A workflow for using nf-core/createpanelrefs for PON generation is in place and documented from a production perspective
Blockers
No response
Anything else?
No response
I think it will be difficult to prioritise this work. At the moment the PON workflow in balsamic is working and is quite simple to maintain as it is sharing nearly every rule and code with the regular balsamic workflow.
This also ensures that the bamfiles and bedfiles used in the production workflow of balsamic is the same that are used to create the PON. In my opinion I think we should first decide on what pipeline we will use in the future, and then once we have decided we can migrate all workflows that we need otherwise it can lead to a bit of extra work...
For instance, if we had migrated the PON workflow before release 16, we would have needed to update both the nfcore PON workflow to be in sync with the UMI-changes and dynamic padding of the bedfiles we added in release 16.
