search

eboileau / ribotools

Miscellaneous programs associated with the rpbp package and more
MIT License
1 stars 0 forks source link

Create count tables for predictions? #1

Closed eboileau closed 1 year ago

eboileau commented 1 year ago

Currently, run-htseq-workflow uses the GTF from the config (or the newly created GTF if de_novo_gtf is present), but type=CDS is used for counting. It is possible to override defaults e.g. using --type exon, but somehow this didn't work as expected (possibly due to concat gtf and de_novo_gtf? Many entries are expected to overlap or even to be duplicated).

But we could create our own GTF using the predictions, and run HTSeq-count on this, which would really give us count tables for the ORFs.

eboileau commented 1 year ago

Added get-gtf-from-predictions, but still needs to be tested on the TEA downstream. Option --gtf added to run-htseq-workflow. We need to add explanations in the docs regarding the usage of GTF with htseq-count --type X, in particular if de_novo_gtf is present.