DeSEQ normalization for knockout

Question

[langya]

Hey

I have two different knockout (knockout of different genes), and one negative controls (all knockout by CRISPR Cas9 SYSTEM). First i normalized my knockout to the negative controls to see what genes are differentially expressed between WT and knockout by using deseq. Now I want to compare what genes are differentially expressed between my knockout, can i take the DeSEQ output directly to compare or what should I do? How does the normalization in DESEQ work? [The normalization to negative control is necessary because I do not want to see the changes that are purely cause by CRISPR system. ]

Answer 1

[Fabio Marroni]

The best option is to feed ALL of your raw counts to DESeq2 and normalize all the data together. DESeq2 (and if I well remember DESeq as well) has options to perform analysis when you have more than two levels (treated and control). You can find some mention of it at the paragraph "Contrasts" of the DESeq2 manual