Entering edit mode
Hi, I am looking to do some downstream analysis on salmon output for non-coding RNA and I would like to look at transcript levels.. I am aware of the option of wasabi --> sleuth, but I was wondering if anyone has tried directly to use IsoformSwitchAnalyzeR or TSIS with transcript counts/abundance?
I have (naturally since I'm the authour) used IsoformSwitchAnalyzeR quite a lot. IsoformSwitchAnalyzeR directly supports data from Salmon (use both abundances and counts as described here) and allows a wide range of analysis for Isoform Switches as well as alternative splicing both on individual genes as well as genome-wide summaries (examples are in this part of the vignette). If you have any questions please fire away :-)
I have a few ;)
First of all, have you looked at non-coding RNA? These are transcripts that are transcribed (hence RNAseq) but not translated. A lot of the downstream analysis in the vignette is about protein analysis, no?
Also, I am still having difficulties with the switchAnalyzeRlist, as you mentioned in the googlegroup
https://groups.google.com/forum/utm_medium=email&utm_source=footer#!msg/isoformswitchanalyzer/4Ypxfd44qYY/rmOerpKbAwAJ
I should try to update Bioconducter to post 3.7, so I am trying to do that... also stuck on that, there are a few packages that I am getting the error that they are not write-able, but they have all the permissions.. so I am looking into this.
There are loads of examples where the switch is between a coding and a non-coding transcript - see fx figure 4 in our article.
Unfortunately apart from the current 3 possibilities: 1) Analysis with CPAT (verifying the coding potential). 2) Looking at structural differences (lengths, number of exons etc) and 3) analysing alternative splicing predicting interesting functional consequences between non-coding RNAs are not something where a lot of tools exists. I have tried with Rfam - but it always find so few motifs (very small percentage of transcripts are annotated with anything) that a integration into IsoformSwitchAnalyzeR have never been worth the while. If you have any ideas please let me know :-)