Hello all, currently I'm working on the de novo assembly of a non-model plant. I've got tissue-specific illumina-, as well as whole-plant ONP-cDNA-reads. My current plans are as follows:
- Conducting master-assemblies (short read & Hybrid) using the three best assemblers out there (Trinity, transabyss, rnaspades, according to Hölzer & Marz 2019)
- Merging these assemblies
- Assembly evaluation (Busco, Transrate, Salmon)
- Quantification
- Analysis of DEG
Regarding the merging of transcriptomes I've got a few questions: I'd like to build supertranscripts using Corset & Lace (Davidson et al. 2017). How does one go about annotating these for later differential expression analysis?
How do you analyse the quality of supertranscripts?
Are other merging techniques better suited? (There are alternative approaches, namely selecting longest sequence per assembled transcript, longest coding region, or by shared alignment [Gilbert 2019, Davidson et al. 2014])
If anyone has some insights into this, I'd be very thankful.
Edit: changed "differential expression" to "differential expression analysis"
Sorry, it's not my particular area, but I will see if I can find somebody.
Thanks for the offer, although in the end I simply conducted several naive assemblies (Trinity, rnaSPAdes & a custom hybrid pipeline for ONP + illumina). I then simply picked the highest quality one (based on technical & biological metrics). Although transcriptome merging is being held as potentially useful, there are just no standard ways of doing it ATM. rnaSPAdes was pretty solid BTW, despite working the high-level polyploid genome (octoploid!).