Interaction Terms in DESeq2 - Shrinkage
1
0
Entering edit mode
@andrebolerbarros-16788
Last seen 9 days ago
Portugal

Hey everyone,

I have a RNASeq dataset, with three variables - Time Point (1,10,100), Drug Concentration (0,0.001,0.1) and Cell Line (LineA, LineB, LineC). Therefore, I have been working with the design: ~Time.Point*Cell.line*Drug.

My question is, according with the interactions section of DESeq2 vignette & example 3 from results function, to get the comparison between 1 and 10 days in Time Point, at the baseline drug concentration (0), for Cell.Line B, I should do:

results(dds, contrast=list( c("Time.Point_10_vs_1","Cell.LineBTime.Point10") ))

However, if i want to get the log2FC after shrinkage, do I need to reorganize the levels (put Cell.LineB as reference level) and then run nbinomWaldTest? Or is there another way to do this?

Thanks in advance

DESeq2 • 239 views
ADD COMMENT
0
Entering edit mode
swbarnes2 ★ 1.4k
@swbarnes2-14086
Last seen 1 hour ago
San Diego

I think the simpler way to do it is to make a new column of conc_time_tissue all stuck together, and make that column your design.

Then you can easily compare one subset of samples to another.

ADD COMMENT
0
Entering edit mode

Thanks for your suggestion!

For apeglm shrinkage, on that scenario, I would have to relevel the variable and then re-run with nbinomWaldTest, correct? That would require similar iterations of relevel and rerun the model as the interaction model, I think.

ADD REPLY
0
Entering edit mode

Yes, apeglm requires relevel and nbinomWaldTest to generate the coefficient estimate and its SE.

Alternatively you can use ashr with a contrast.

ADD REPLY

Login before adding your answer.

Traffic: 388 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6