the final assembly is produced by the EVIGENE_TR2AACDS module. In the denovotranscript.nf workflow, I have created ch_transcripts which contains the fasta file with rna sequences - this is probably what you want to use for transdecoder edit: I just realised I will need to look into how Evidential Gene deals with RNA that's not mRNA. I've mainly been focusing on mRNA, but I noticed some of the annotation steps apply to other types too

Is the type of studied RNA defined in the samplesheet per row? Can it be both, at the same time? If so, the pipeline could then decide which ones to run depending on that. We can always have extra parameters to specifically let the user choose assembler, but we can guide them with the default values

It's not defined in the samplesheet, but that's a great idea. Right now, we pool reads from all the samples to make the assembly, so all the samples would need to be analysed with the same RNA type - this would be a common scenario, right? You may need to edit that if you would like to give people the option to choose different RNA types. Alternatively, we assume that if they want different RNA types, they would just re-run the pipeline with different params/samplesheet values

We could do that, yes! @Eleftherios Bochalis what do you think? Do you usually have mixed samples or is it only mRNA?

The main idea I had, was that if someone has performed bulk RNA-seq on samples that will be subjected for de-novo assembly, then all RNA-types might be useful, hence the Infernal, tRNAscan-SE steps. I did not know that EVIGENE_TR2AACDS drops ncRNAs. We can either split this into two steps (mRNA goes through EVIGENE_TR2AACDS, ncRNAs go directly to Infernal) or we can try to find other tools meant for transcriptome merging that are more permissive to ncRNAs

sure, both could work. I'm leaning towards this option:

split this into two steps (mRNA goes through EVIGENE_TR2AACDS, ncRNAs go directly to Infernal)

if you decide to go with other option, you may need to do a bit more research to find a tool? I had chosen Evidential Gene due to the comparison it makes here: https://www.biorxiv.org/content/10.1101/829184v1

This option is definitely the more straightforward one. We will check out the paper you shared! Thank you Avani!