Sorry, my question might be so naive.
I want to use the pairedGSEA package to run a paired differential gene expression and splicing analysis. I have already used DESeq2 for differential gene expression analysis. I aligned my reads with HISAT2 and made the count files with the HTSeq count.
In my DESeq data set object, the row names are ENSG IDs.
But I saw in the documentation of pairedGSEA that the row names should have the format 'gene:transcript'.
Should I use a specific tool for alignment and making count files, or how can I fix it?
I really appreciate any help you can provide.
Thank you!
James is exactly right. Both salmon and kallisto can produce transcript counts, if you map to a transcript-level reference. You will need a matrix with gene:transcript names, and tximport can help extracting the count matrix, as suggested.
For inspiration, here's how I extracted transcript counts from ARCHS4 for the paper.
If you want to do a splicing analysis, you cannot have counts/gene because the counts are aggregated at the gene level, not the transcript level. You could use the salmon aligner to get transcript counts and then you could either simply use the quant.sf files directly (modifying so you have gene:transcript identifiers), or maybe you could use tximport, although I believe that will length scale the transcript counts, which you may not want? I am not familiar with pairedGSEA so you will need to peruse the vignette for both that package and tximport to make sure.
Hi Sara,
Apologies for the slow response!
James is exactly right. Both
salmonandkallistocan produce transcript counts, if you map to a transcript-level reference. You will need a matrix with gene:transcript names, andtximportcan help extracting the count matrix, as suggested. For inspiration, here's how I extracted transcript counts from ARCHS4 for the paper.Hope you got/get it up and running.
Best regards,
Søren
Developer of
pairedGSEA