On short reads, Kallisto/Salmon is a standard workflow for measuring RNA transcript counts in TPM. However, when I tried to Google a similar workflow for long reads, it's not clear there is a replacement.
$\begingroup$
$\endgroup$
Add a comment
|
3 Answers
$\begingroup$
$\endgroup$
While the selective-alignment algorithm that is part of salmon is designed for short reads, salmon also supports quantification via alignments in a SAM / BAM file (aligned by other tools). You have to be somewhat careful about setting the proper command line flags, but nonetheless, the quantification should work fine. You can, for example, find a pipeline that uses minimap2 for alignment followed by salmon for quantification of long reads here.
$\begingroup$
$\endgroup$
Stringtie now does long reads too. We've found flair to be quite difficult to install and to crash often. That said, stringtie isn't very good about showing you its working (e.g. I have no idea how to get it to give the assembly it uses to figure out all the isoforms). Its values for TPM/FPKM also seem a little weird, in that they're perfectly correlated with each other, and for transcripts they're perfectly correlated with coverage too.
https://genomebiology.biomedcentral.com/articles/10.1186/s13059-019-1910-1 https://github.com/gpertea/stringtie
TALON is another tool that does isoform assembly. I'm not sure whether it does quantification though. https://github.com/mortazavilab/TALON#
$\begingroup$
$\endgroup$
https://github.com/BrooksLabUCSC/flair was excellent. Highly recommended. One could also ask for TPM.
answered Aug 26, 2020 at 0:11