limma-voom multiple factors & groups

Hi Galaxy Community,

I am trying to use the limma-voom tool and I have been following the “Genes to Counts” training. Within the training it gives the tip about “multiple factors” but the tip does not give an example of how the input would change when entering the “contrast of interest”. I have tried various different inputs to try and figure out on my own but when I do, I always get an error and from the errors it seems that the program will not read pass the 2nd column of the either the factor file or when inputted manually.

Here is what my factor file looks like:

What I am trying to do here is take all of the samples from the “Reaction” factor that are in group “A” and contrast the “Treatment” factor of groups “T-P”. I continue to receive an error stating that object “P” cannot be found which leads me to believe that the program is only reading the 1st column (sample IDs) and the 2nd column (the “Reaction” factor. I tried entering the multiple factors and groups manually but either way on the first factor and groups are accessed.

Can someone please assist me with either how to input the “contrast of interest” to account for multiple factors and groups or can some explain it to me differently (maybe I am missing something here)?

1 Like

Hi @montgopw

Correct, the contrasts string can only contain one - per grouping. Why? Those are interpreted.

You could input two different lines, or two different blocks on the form. Or create a compound term with parenthesis. The form has examples of this right under each input area, plus a link to the Bioconductor manual (these terms are passed directly to the underlying tool, so the formatting is identical).

And, maybe your example was simplified … but is there a reason why all treatment groups couldn’t have the same designation for this run? Meaning, the second column of your file could have just T for all. That would simplify your input. If you are using collections with group tags, it would be pretty quick to relabel the group tags for a collection that way (See Hands-on: Group tags for complex experimental designs / Using Galaxy and Managing your Data).

Let me know if I am misunderstanding. And, I’m guessing that you have seen this simple example already, but I’m linking it in for others reading → Hands-on: 2: RNA-seq counts to genes / Transcriptomics