Multifactor Design For RNA-seq Analysis

1

Hello everyone,

I'm performing a RNA-seq analysis using DESeq2 and I'm wondering what is the correct analysis approach for my multifactor design. We have searched different posts in different forums and I can't figure out how to approach it.

Here I paste my multifactor design:

enter image description here

Using a ~batch + condition design, we obtained the following error:

some variables in design formula are characters, converting to factorsError in checkFullRank(modelMatrix) :
the model matrix is not full rank, so the model cannot be fit as specified.
One or more variables or interaction terms in the design formula are linear
combinations of the others and must be removed.

We have also checked DESeq2 multiple factors nested design and we are still obtaining the same error. What is the correct factor design for this experiment?

Thank you so much for advance,
David.


Rna-seq


batch


correction


DESeq2


multifactor

• 279 views

The problem here is that because experiments all used the same sequencing design ("fr_firststand", "fr_secondstrand" or "unstranded"), it is not possible to disentangle what is a library strategy specific effect, and what is an specific experiment effect. You thus cannot correct for both.

No matter though, its almost certainly not neccessary to correct for both.

If you correct for exp, but not for batch, the exp correction factors will encompass the library design specific effects. Thus the design ~condition + exp will correct for any different in library construction strategy.

However, correcting for exp involves computing 4 coefficients, which will reduce accuracy/power. If the only thing different between experiments is the library construction strategy, you might consider correcting for "batch" and not "exp".


Login
before adding your answer.

Traffic: 2239 users visited in the last hour



Source link