cuffdiff vs cuffnorm vs cuffquant

Question

[vaughandy]

Hi all,

I'm new to this so bear with me. I've been using Galaxy to process raw RNA-Seq FASTQ data into FPKM matrices. My understanding is that you can't directly compare the cufflinks output of different samples, that rather you have to use some comparison algorithm to be able to appropriately compare.

I've gone directly from cufflinks to cuffdiff and that seems to give me a perfectly useable gene_FPKM tracking file. However, is that appropriate? Or do I need to use cuffnorm or cuffquant at some point?

I appreciate your help whoever has a spare minute!

Answer 1

[Devon Ryan]

cuffdiff takes care of the cuffnorm and cuffquant portions of the analysis for you, so going from cufflinks to cuffdiff (presumably with cuffmerge in between) makes sense.

As an aside, I would encourage you to not use cufflinks/cuffdiff without a good reason. Unless you really need to find novel isoforms you'll be better served with other tools. In particular, cufflinks is seriously out-dated and has been replaced by stringTie.

[Malcolm.Cook]

Hi Devon,

What in you view is contemporary update on cuffdiff's Differential splicing tests

or does that still count as a "good reason"

Thx!

[Devon Ryan]

I'd prefer to use salmon or kallisto followed by edger or a similar tool, they're inevitably easier to debug.

Answer 2

[vaughandy]

Thanks! I used cufflinks for historical reasons, but I'll give stringTie a shot!