Question

Deseq2. Cluster of an specific list

0

Entering edit mode

chariko • 0

@chariko-8827

Last seen 8.6 years ago

Spain

Hi,

I'm using DESeq2 for DE analysis.

I followed the manual guide for DESeq2 and I was able to perform a cluster taking into account the list of the most significant genes. I would like to do also a cluster taking into account a list of specific genes. Does anybody know how to select a specific gene list?

Thanks a lot for your help!

deseq2 heatmap clustering file list • 4.3k views

ADD COMMENT • link updated 8.6 years ago by Michael Love 41k • written 8.6 years ago by chariko • 0

0

Entering edit mode

It would be helpful if you could give us more details. Can you post the code you used to cluster the most significant, and can you specify the format of your list of genes you want to cluster?

ADD REPLY • link 8.6 years ago Michael Love 41k

0

Entering edit mode

The code I used for generating the most significant genes was the following:

res <- results(dds)
resOrdered <- res[order(res$padj),]
select_genes<-rownames(subset(resOrdered, padj < 0.1))

rld <- rlog(dds)
heatmap.2(assay(rld)[ select_genes,] ,density.info = "none",symm=F,Colv =T,trace="none", col = colorRampPalette(c("red", "green")))

For selecting the gene list, first I downloaded the significant genes with the following code:

write.csv (as.data.frame(resOrdered),file="results_Deseq2.csv")

Of this "results_Deseq2.csv" file, I selected a list of 154 genes which I would like to make a cluster with. How can I do this?

ADD REPLY • link 8.6 years ago chariko • 0

score 2 · Accepted Answer · 2015-09-24

So this line:

assay(rld)[select_genes,]

is the one where you subset the rlog data values. assay(rld) is a matrix with genes as rows and samples as columns.

take a look at:

head(select_genes)

and you can see how assay(rld) can be indexed to subset certain genes by name. You just need to provide a vector of gene names in the same way, by defining a vector of gene names your_genes and then using the code:

assay(rld)[your_genes,]