How do i perform DGE if based on PCA the samples are separated along PC2 and not PC1?

Question

[druggable]

Hi All,

I have gene expression data for multiple groups and need to perform differential expression analysis. Results from PCA showed that the experimental groups can only be distinguished along PC2 (18% variance). Do I need to adjust my data to account for PC1 (32% variance)? Can you point me to a general workflow how to do this?

Thanks!

Answer 1

[swbarnes2]

If you want, you can make the PC1 values another column of sample information, and use it as another element in the design. But it might be better if you can figure out what PC1 comes from (batch effect? sample gender? ) and use that as another element of the design.