DESEQ2 design matrix help

[adrian]

Hello:

I have a complicated RNASeq experimental design and I need some help from experts.

my aim is to compare normal vs. tumor rna-seq data.

I have 4 different normal samples obtained from two different sites of two different animals.

Similarly I have 9 tumor samples. 6 of them were obtained from 3 different animals (from 2 sites). 3 samples from 3 different sites of a single mouse.

Normal
N1 - animal 1
N2 - Animal 1

N3 - Animal 2
N4 - Animal 2

Tumor:

T1 - Animal 1
T2 - Animal 1

T3 - Animal 2
T4 - Animal 2

T5 - Animal 3
T6 - Animal 3

T7 - animal 4
T8 - animal 4
T9 - animal 4


Using DESeq2, how can this model be fitted for identifying differentially expressed genes between normal and tumors.

thanks

[ayushraman]

Conditions = (Normal and Tumor)
Genotype = (Animal1, Animal2, Animal3, Animal4)
DESeq equation ~ Conditions + Genotype

One concern/question is how do you want the comparison to be ?

Comparison 1: Normal_Animal1 vs Tumor_Animal1
Comparison 2: Normal_Animal2 vs Tumor_Animal2
Comparison 3: Normal_Animal1+Normal_Animal2 vs Tumor_Animal1+Tumor_Animal2 (I am not sure if this is correct)
Comparison 4: Normal_Animal1+Normal_Animal2 vs Tumor_Animal1+Tumor_Animal2+Tumor_Animal3+Tumor_Animal4 (I am very skeptical about this comparison because you are not comparing apples to apples..)