To do two kind of variables to do differential expression analysis

Question

[1106518271]

A toy example:

health mouse: tissue a(replication 3), tissue b(replication 9).
disease mouse: tissue a(replication 5), tissue b(replication 6).

For health_a_1, health_a_2, health_a_3, health_b_1...each do RNA-seq, got their RPKM list

I hope to see tissue a has differences in expression between health and disease, tissue b has differences in expression between health and disease.

How to know there any differences in expression? I know one factor, but here two: health condition and tissue. I think here shouldn't study independent, use model y=ax1 + bx2 + cx1x2?

Some suggestion? Thanks!

[Benn]

First off, RPKM values are not suitable for sound statistics. They (all bioinformaticians) recommend to use raw read counts. When you have raw read counts you can continue for example in limma (voom or trend) to make designs like yours. Read the manual, it is pretty well written (also for beginners).

[1106518271]

Thanks! What's more, for id list (my row of matrix), recommend use mRNA id expresion or gene id expression in general?

[Benn]

Try to use something like featureCounts, with gene annotation from ENSEMBL. Then you'll have ENSEMBL gene names, which you can convert later to anything you want.