Unsupervised hierarchical clustering of DEG in Rstudio

Question

[MS]

Hi,

I want to perform unsupervised hierarchical clustering of DEG (output of edgeR) using Rstudio.

I found hclust and dist functions but not sure about which packages should I use.

I checked posts but that confused my mind more.

Could you please explain me the correct tools and methods for it?

[liorglic]

What do you mean by "clustering of DEG"? Based on what? what would be the distance measure between DEGs?

[MS]

Based on the expression profiles of differentially expressed genes obtained using EdgeR pakcage.

[liorglic]

Maybe I'm missing something, but I still don't understand. Maybe if you elaborate on your experiment design and expected output, then I (or someone else) could help. In what case would you say two DEGs should be clustered together? Are you trying to perform some kind of co-expression analysis across conditions? Can you provide a link to a paper where this type on analysis was performed?

[MS]

Sorry for the unclear explanation.

I have 2 groups as normal and patient and I have DEG list obtained with comparing these groups.

I just want to get heatmap showing expression profiles like the paper given the link below

http://cancerinflammation.dk/article/Diagnostic%20miRNA%20profiling%20in%20CTCL.pdf

[liorglic]

No problem - got you now. So you want to cluster DEGs and samples. You need to create a matrix where the cell in row i and column j corresponds to the expression level for DEG i in sample j. Once you have that, I think this is what you need.

Answer 1

[jared.andrews07]

You'll likely find the dittoSeq dittoHeatmap() function of interest. It can make heatmaps from your DGEList object directly for whatever genes you specify (and it's color blind-friendly by default).