This is probably a simple fix but I am trying to run roast using curated gene sets from the MSigDB. I have been able to run roast successfully in the past using GO gene sets but am struggling to use these other type of gene sets.

I am getting an error message when I try to create the list of gene sets I'm exploring . The four bolded lines are where I'm pretty sure I'm making an mistake. 

summarized.counts <- read.table("C:/Users/Alicia/Documents/summarized.counts.matrix", row.names=1, header=TRUE, sep="\t")
geneset <- read.csv("C:/Users/Alicia/Desktop/TEST.txt", row.names=1, sep="", stringsAsFactors=FALSE)
design_two<-model.matrix(~0+ factor(c(1,1,1,2,2,2,3,3,3)))
colnames(design_two)<-c("stage1", "stage2", "stage3")
con<-makeContrasts(stage3-stage1, levels=design_two)
library(org.Hs.eg.db)
importgene <- c("M5539", "M2222")
importgene <- unique(importgene)
Rkeys(org.Hs.egGO2ALLEGS)<-importgene
 

How exactly to a create a list of those gene sets of interest and what data base do I then use to associate that information with my sequence?

I'm having a hard time relating the code you provided with the idea of using curated gene sets from MSigDB.

There are many ways to get MSigDB genesets into a form that you can use inside the R/Bioc universe. Gordon's lab provides *.Rdata files for each of the collections (for human or mouse) that you can use directly, if you like, which you can download from this web page.

Also, this part of your code:

library(org.Hs.eg.db)
importgene <- c("M5539", "M2222")
importgene <- unique(importgene)
Rkeys(org.Hs.egGO2ALLEGS) <- importgene

I can't parse much, at all. What are you expecting to happen when you use MSigDB specific identifiers (M5539 and M2222) and assign them to Rkeys(org.Hs.egGO2ALLEGS)? Is there some code example somewhere that you can link to that suggests this way of doing something?

Lastly, I suspect the Rkeys stuff is from an older way of interfacing with these database-like things, and you're now encouraged to interact with these db-like resources using the select interface. You can read more about them in the AnnotationDBI vignettes, which are accessible via the AnnotationDBI package homepage.

I apologize, I realize I wasn't making any sense. When I originally ran roast I created a list of GO gene sets I was interested in. This was what I was doing on the line importgene<- c("...."). Originally those had GO IDs, however I changed them to see if it was possible to run curated gene sets instead (M#). I then had to associate those gene sets with the genes in each and my sequence. I guess I just don't understand how to run roast, while only looking at 10 or so curated gene sets of interest.

I apologize, I am new to R and gene set analysis.