As a clinician I want to be able to "verify" if the OTUs the classifiers have identified are correct so I can feel certain in my interpretation.
Suggested approach
A click button to run post-processing in a separate repository to validate the interesting species.
KrakenTools will be used to extract reads from Kraken2 report. Custom scripts will be needed to extract reads from reports generated by other classifiers.
Run BLAST separately for the species of interest.
Mapping: Map the extracted reads against the target genome and the human genome.
Can be closed when
[ ] We've decided on what a good approach to this is
Blockers
No response
Anything else?
We have to figure out how this should be approached. There are multiple things to take into consideration.
If the visualisation tool should launch new compute jobs, this needs to be agnostic to infrastructure as this is a national tool.
If the visualiser starts performing calculations, this might push it into the frame of being a IVDR classified product and would have to comply to those rules, which are extensive when it comes to products with a GUI.
Could the pipeline perform these alignments, and if so, on which OTUs? Can't be all of them.
Need
As a clinician I want to be able to "verify" if the OTUs the classifiers have identified are correct so I can feel certain in my interpretation.
Suggested approach
Can be closed when
Blockers
No response
Anything else?
We have to figure out how this should be approached. There are multiple things to take into consideration.