This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Extracting Gene Names From Venn Diagram

I created a venn diagram (5 intersecting circles) using the code below, but not sure how to go about extracting the genes that are differentially expressed across different compartments.

## create experimental design
experimentalFactors<-targets$time
sampleGroups <- factor(experimentalFactors, levels = unique(experimentalFactors))
exptDesign <- model.matrix(~ 0 + sampleGroups)
colnames(exptDesign) <- levels(sampleGroups)

## create contrasts
contrasts <- makeContrasts("hr6-0","hr24-hr0","hr48-hr0","hr72-hr0","hr80-hr0",levels = exptDesign)

fit <- lmFit(selDataMatrix, exptDesign)
fit2 <- contrasts.fit(fit, contrasts)
fit2 <- eBayes(fit2)

## venn diagram
fit2.venn <- fit2[,1:5]
results <- decideTests(fit2.venn, method="separate", adjust.method="BH", p.value=0.05, lfc=1.5)
vennDiagram(results)

Please help!

r limma bioconductor microarray
• 8,123 views • link
updated by Irsan Irsan • written by robjohn7000 robjohn7000

1 answer

The results-object you have that was returned by decideTests() is a TestResults-class. This is basically a matrix with rows as genes and colums as contrasts. Each value is -1, 0 or 1 meaning downregulated, unchanged and upregulated. So if you want to get all genes that are different in contrast 1 but not in 4 you can do

yourGenes<-which(results[,1]!=0 & results[,4]==0)
results[yourGenes,]
• 158 views • link
written by Irsan Irsan

Brilliant! That worked nicely. Thanks Irsan for the code and the explanations. I was able to get the genes that are upregulated in all contrasts with the following:

         Genes <- which(results[,1]==1 & results[,2]==1 & results[,3]==1 & results[,4]==1 & results[,5]==1)
         Genes <- results[Genes,]

But the gene names are in the "illumina" format and I wanted the official gene syambols, and then tried two ways (neither worked):

         Official_genes <- fit2$genes[Genes, "geneSymbol"]


         Official_genes <- fit2$genes$geneSymbol[Genes]

Is there a way of getting the genes in the official gene symbol format?

• 0 views • link
written by robjohn7000 robjohn7000

Well that depends on whether the official gene symbols are somewhere in your data in fit2 in the first place. Is it?

• 132 views • link
written by Irsan Irsan

yes, it is here:fit2$genes$geneSymbol

• 0 views • link
written by robjohn7000 robjohn7000

I guess fit2$genes is a data frame. Does it also contain a column with the illumina ids you mentioned earlier?

• 0 views • link
written by Irsan Irsan

yes it's the same data frame that contains a column of IDs.

• 0 views • link
written by robjohn7000 robjohn7000 
Genes <- names(Genes)
Official_genes <- fit2$genes[Genes, "geneSymbol"]
• 340 views • link
written by Irsan Irsan

Many thanks, your latest code worked with a modification, thus: Genes <- row.names(Genes) Official_genes <- fit2$genes[Genes, "geneSymbol"]

I really appreciate your help Irsan!!!

• 218 views • link
written by robjohn7000 robjohn7000

You are welcome :-)

• 0 views • link
written by Irsan Irsan

Log in to answer this question.

More posts like this