Hello!!!
I’m running Deseq2 to find differentially expressed genes in breast cancer samples according to genetic ancestry. For this I categorized my samples in three groups according to European ancestry fraction  (G1: 25-50% (15 samples), G2: 50-75% (35 samples), G3 (6 samples): 75-100%). Running these comparisons I found some genes the problem is that when I create just two groups, none gene appeared, why this could happen? I could be related with the dispersion within the groups? How can I analyze this dispersion?

I would really appreciate your help!!

Thank you

Silvia

We typically recommend that you put all the samples into one DESeqDataSet and run DESeq() on this. There is more information in a FAQ in the vignette about this question.

But it's important to recognize that by changing the samples that are being analyzed at once you change the parameter estimates and therefore the p-values.