Search
Question: Problem in genelist levels when running topGO.
0
gravatar for peirinl
8 weeks ago by
peirinl0
peirinl0 wrote:

I am running multi-level experiment of a dataset using LIMMA,

contrast_matrix_Grade<- makeContrasts(AdvanceVsEarlyinT = A.T-nA.T,

                                      AdvanceVsEarlyinnT = A.nT-nA.nT, 
                                      TVsnTinAdvance = A.T-A.nT, TVsnTinEarly = nA.T-nA.nT,
                                      levels = design_Grade)
fit <- eBayes(contrasts.fit(lmFit(eset,design = design_Grade,block = i,
                                                correlation = corfit_Grade$consensus),
                                          contrast_matrix_Grade))
table_fit <- topTable(fit, number = Inf)

results <- decideTests(fit)

 

Since I want to see the gene that simultaneously differentially expressed in two groups, 

intersection <- results[,"TVsnTinHigh"] & results[,"TVsnTinLow"]
table_Grade <- topTable(fit [intersection,], n=Inf)

DE_genes_Grade <- subset(table_Grade, adj.P.Val < 0.1)$PROBEID

back_genes_idx <- genefinder(eset, as.character(DE_genes_Grade), 
                             method="manhattan", scale="none")
back_genes_idx <- sapply(back_genes_idx, function(x)x$indices)
back_genes <-featureNames(eset)[back_genes_idx]
back_genes <- setdiff(back_genes, DE_genes_Grade)
intersect(back_genes, DE_genes_Grade)
length(back_genes)

gene_IDs <- rownames(table_Grade)
in_universe <- gene_IDs %in% c(DE_genes_Grade, back_genes)
inSelection <-  gene_IDs %in% DE_genes_Grade 
all_genes <- factor(as.integer(inSelection[in_universe]))
names(all_genes) <- gene_IDs[in_universe]

ont <- "BP"
top_Grade_GO_data <- new("topGOdata", ontology = ont, allGenes = all_genes, nodeSize = 10,
                         annot=annFUN.db, affyLib = "hugene10sttranscriptcluster.db")

Error in .local(.Object, ...) : allGenes must be a factor with 2 levels

 

> str(all_genes)

 Factor w/ 1 level "1": 1 1 1 1 1 1 1 1 1 1 ...
 - attr(*, "names")= chr [1:2373] "7924029" "7908072" "8008237" "8037408" ...

But when I used table_fit, there is no problem. 

> str(all_genes)
 Factor w/ 2 levels "0","1": 2 2 2 2 2 2 2 2 2 2 ...
 - attr(*, "names")= chr [1:17425] "7924029" "7908072" "8008237" "8037408" ...

So if I only want to see the GO of the intersected genes of both groups, what should I do?

Thanks

 

ADD COMMENTlink modified 8 weeks ago • written 8 weeks ago by peirinl0
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 166 users visited in the last hour