Way to do differential gene expression analysis without writing Rscripts?

Question

[jpcchoy]

Hi, everyone.

I have generated gene count tables from Ribo-Seq data across four different treatment conditions using featureCounts, and would like to work out differential gene expression across the four conditions. I had hoped to use DESeq2, but I've found it difficult as I am not sufficiently competent in R yet. Though I will continue to learn R, I'd like to be able to work out my next round of experiments before long. Is there a quicker way to conduct such an analysis, just on the command line perhaps, so I don't have to wait for my R competency to catch up?

Thanks for any help.

[jared.andrews07]

What in particular have you been having issues with? DESeq2 is pretty user friendly and your design doesn't seem too complicated. The vignettes are quite good, and there are lots of step by step tutorials out there (like this one and this one).

[Benn]

I am not sure if galaxy is still functional, it uses R scripts behind the scenes, but offers GUI interface for the user.

Answer 1

[Charles Warden]

To be honest, I think it would be best if you gradually expand your skillsets, especially since I think you probably should be doing some custom analysis.

There are commercially available programs that don't require coding, but I have shifted away from those over the course of years. In other words, my concern is that something that might seem like it saves you weeks or months of times may actually be a hindrance and/or cause problems on the scale of years. So, even if takes extra weeks or months (for a given round of analysis), I think waiting to learn R is the better option. Plus, you need to have some level of understanding in terms of where the result come from. For example, there are mulitple ways to calculate p-values within either edgeR or DESeq2.

If there was something that started with a graphical user interface and whose goal was to teach you to code (and always provided the R code for publications, to make your results reproducible), then that is what I would recommend (and I would be very happy to hear if that already exists).

Otherwise, I think there are sometimes GUIs for R packages (kind of like Glimma or EDGE for microarray data):

• Benn is right that there are edgeR and DESeq2 implementations in Galaxy. You can find them using the "search" option, although I haven't tested them.
• Again, I haven't tried it, but I believe ARMOR runs R code through a command-line interface
• Yet again, I haven't tried them, but I believe IRIS-EDA, DEBrowser, and iDEP provide a web interfaces to run differential expression
• Yet yet again, I haven't tried it, but I believe iGEAK provides a graphical interface to run R functions