What is the best way to identify which group of genes is more highly expressed?

Question

[pl23]

Hello,

I am new to bioinformatics. I am trying to compare two sets of genes that belong to different subgenomes of my species to identify if one of the subgenomes is more highly expressed. I have gene expression data in the form of counts, with multiple tissue types and replicates. I tried using the usual ANOVA/t-tests using my tissue type and subgenome as factors but my data is highly skewed and does not meet the normality condition, even after applying the log transformation. Additionally, running non parametric Wilcoxon tests was not helpful as my data is not symmetric.

Is there a better way to approach this problem? I am happy to do the analysis for each tissue type separately if necessary.

Any feedback is much appreciated.

Thank you!

Answer 1

[cfos4698]

Assuming these are RNAseq counts, there are many possible pipelines to process your data. Two of the more popular routes are using:

• DESeq2: see this vignette
• Limma+voom: see this vignette

These approaches will expect raw, untransformed counts. And it's expected that the counts don't follow a normal distribution.

[pl23]

Hi, Thanks for your response! Yes I already considered DESeq2 - unfortunately I think it can only accommodate tissue-wise or treatment-wise comparisons, not genewise comparisons. Would Limma/Voom be able to do this?