Heatmap: edgeR counts and DEG

Question

[RiNG]

I have used edgeR in Galaxy to perform differential expression analysis. As output I have 3 files with a list of differentially expressed genes (comparison between 3 different groups of samples) with log(FC), log(CPM), FDR, etc. I also have a separate file containing the normalized counts, but for all the different samples within each group.

To make a heatmap out of the diferentially expressed genes, how can I cross the information between the counts file and the DEG files to select only the genes of interest in the counts file?

Thanks in advance.

[Vasu]

Take your Differential expressed genes list and get the counts data of those genes. Convert counts to logCPM and use that for heatmap.

[Kevin Blighe]

It would really help us if you pasted samples of the data that you have. Otherwise, we can only speculate as to their formatting / structure.

[RiNG]

How can I get the counts data for specific genes in an Excel file with >20000 genes? If I use the "Find" function I would take ages.

[Kevin Blighe]

How did your data end-up in Excel? Better to export your Excel data in TSV or CSV format, and then read that into a R Programming Language environment.

[RiNG]

It was a TSV file I got as output in Galaxy after running edgeR.

If I open this file in Rstudio, how can I then find the genes I am interested in?

[Kevin Blighe]

Can you paste the top-left corner of the data? I assume that it is genes as rows and samples as columns?

[RiNG]

GeneID 19_c2.trimmed.fastq.sorted.bam 24_c3.trimmed.fastq.sorted.bam .....

ENSSSCG000000...

ENSSSCG0000...

This is it; so yes rows is genes and columns for samples (8 samples).

[Kevin Blighe]

These are ENSEMBL gene ideas Sus scrofa (pig). You can likely convert these using biomaRt package in R. Galaxy should also have a gene conversion tool, no?

[RiNG]

That is not a problem. My problem is that I have +20000 genes in a counts file and I only want to select a few.

Is there a function in R that can find and return only the rows I am interested in? And reduce the counts matrix from 20000 to 100 genes of interest?

[Kevin Blighe]

Yes, if these ENSEMBL IDs are the rownames of your object, then just do this:

genesOfInterest <- c("ENSSSCG0000001", "ENSSSCG0000056", "ENSSSCG005555", "ENSSSCG000009", "ENSSSCG003332")

MyData[which(rownames(MyData) %in% genesOfInterest),]

There are a few ways of doing it, though.

[RiNG]

It works! Thank you for your time.