Compare gene expression in mice and human RNA-seq data

Question

[Wang Cong]

Hi,

I have samples from mice and human, treated with drugs of similar effect. I can do differential expression analysis for mice and human separately. I wish to compare the gene expression between mice and human. I think there are several ways but I not sure if they are correct:

1. choose some genes of interest and compare the orthologs between mice and human (does comparing the normalized counts, e.g. TPM, FPKM make sense)?

2. do correlation analysis (e.g. pearson) on a set of gene (e.g. all upregulated genes (log2FC>1 padj<0.05), all gene of interest...)

3. compare the enrichment analysis results between mice and human to see if there are similar enrichment.

4. combine the read count matrix of mice and human and perform DEG analysis together (I think this is complicated and not sure if it makes sense)

[ATpoint]

1) To me this makes no sense. RNA-seq is relative, not absolute so you cannot compare the values.

2) Will give you a sense of the data.

3) Yes, that's what I'd do. Take either human and mouse as reference, use the DEGs to make genesets and see how this looks in GSEA in the opther species. A combination of 2+3 (sunset to relavant genes) should give an idea about species conservation effects.