Question

normalization of RNA-seq data after over-expressing the transcription factors

1

Entering edit mode

Bogdan ▴ 670

@bogdan-2367

Last seen 13 months ago

Palo Alto, CA, USA

Dear all,

talking about the standard normalization methods of RNA-seq data as implemented in limma, edgeR or DeSeq2 :

is there any modification we shall do in the normalization method when analyzing the RNA-seq data of these 3 sets of samples :

a) control b) over-expressing X protein (Wild Type) WT b) over-expressing X protein (Mutant) MUT

taking into consideration the fact that XproteinWT may have very different levels of expression that XproteinMUTANT

(i.e. 1.5-2x fold difference) ?

thanks a lot,

-- bogdan

limma deseq2 voom edgeR • 1.1k views

ADD COMMENT • link updated 5.4 years ago by Aaron Lun ★ 28k • written 5.4 years ago by Bogdan ▴ 670

0

Entering edit mode

I must say that I don't understand your question at all. What exactly is the problem here?

ADD REPLY • link 5.4 years ago Aaron Lun ★ 28k

0

Entering edit mode

Hi Aaron, great to hear from you. If I may re-phrase the question please :

considering the RNA-seq for a set of 3 treatments :

(a) control
(b) over-expressing X protein (Wild Type) i.e. X_WT
(c) over-expressing X protein (Mutant) i.e. X_MUT

I have used limma and edgeR for normalization and differential expression. So far, so good ...

However some colleagues mentioned that, :

as the expression of X_WT is 2*higher than X_MUT (after over-expressing these proteins), the "standard" normalization pipelines shall be adjusted / changed to include this fact.

The question would be : is this a legitimate point ?

if the answer may be yes, the next question is : how shall we include the differences in the expression of XWT and XMUT in the pipelines ?

(we do not have SPIKE-IN controls in the bulk RNA-seq data) ;

thank you !

bogdan

ADD REPLY • link 5.4 years ago Bogdan ▴ 670

score 3 · Accepted Answer · 2019-07-23

3

Entering edit mode

Aaron Lun ★ 28k

@alun

Last seen 19 hours ago

The city by the bay

is this a legitimate point ?

No. I can only assume that something was lost in translation because that doesn't make sense to me. The relative levels of X - in and of themselves - have no bearing on the normalization for other genes. (You would only encounter problems if the change in expression of X caused a majority of other genes to be DE, but that would be rare.)