I have a general quesiton concerning the described workflow of "rnaseqDTU" https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6178912.3/pdf/f1000research-7-17982.pdf

I have followed the workflow to test for DTU on count data generated by salmon. Just as described in the workflow, I have imported them with the recommended packages and applied DRIMSeq filtering and used DEXSeq for transcript counts instead of exons. In the "normal" workflow of DEXSeq when applying to exon counts, there is a function called estimateExonFoldChange to estimate the fold change. I was wondering if this function, if applied to a resulting DEXSeq object that was generated using transcripts instead of exons for the analysis, can be interpreted as intended. I might be confused with the general interpretation of the function's result. Id be glad for help

Im trying to reproduce the results found in one cohort in another cohort and since I cant find the same DTUs I wanted to have some measure to see if they at least agree (on direction of the change for example).

The interpretation should be similar, although there may be fewer transcript features compared to exon features, when the fold changes are calculated. But roughly, it's the fold change for counts from "this" feature over the counts from "other" features.