Closed dlaehnemann closed 9 months ago
Maybe we should also refactor the write_results()
function quite a bit. This is a big monolith, that can probably do with a bit of cleanup. But I'll first add some proper QuantSeq testing data to the fix-canonical...
branch to get the test suite passing again.
The underlying problem that we identified with the work on this
debug-vroom
branch was a malformattedcustom
file for specifying canonical transcript to use insleuth-diffexp.R
. The take-away message here were:sleuth-diffexp.R
script with its largewrite_results()
function was hard to debug, and to ease the burden a bit in the future, we should probably keep the extra logging statements we included.datavzrd/diffexp-template.yaml
does not seem to play nice withcustom
canonical transcript files. Thecanonical
column does not make sense in thegenes_aggregated
results table (as this should only contain gene names / identifiers, and no transcript identifiers, and only the latter could be canonical or not), so we remove it there. However, how to have acanonical
column in thegenes_representative
case with acustom
canonical transcript file still needs to be solved before merging this PR.