detecting predominant exons using DEXSeq, edgeR

[alpesh]

Hi all,

I have RNA-seq data for 3 conditions (with biological replicates), and I`m hoping to investigate how the exon usage varies across conditions. I am using testGeneForDEU and spliceVariants in DEXSeq and edgeR repsectively to identify genes that show evidence of differential exon usage. I want to go a step further and compare exonic expression across conditions. For example, I want to see if behavior of a gene for condition 1 is similar to condition 2 or 3.

So if an exon is expressed more in condition 2, and little in condition 3, which way does condition 1 behave, and maybe classify it as 2-like, 3-like or mean of 2 and 3? Is there a way to statistically look at this? Both packages provide nice way to visualize this through plotexonUsage and plotDEXSeq.

Thanks,
Al

[areyes]

Hi @alpesh,

I think you could use the fitted "splicing" coefficients from DEXSeq to do this (e.g. what it is plotted in plotDEXSeq), you could get them with the function estimatelog2FoldChanges, and change the parameter getOnlyEffects to TRUE. I guess what you could do is some clustering of exon usage profiles and look at similarities in exon usage profiles.

Alejandro