Interaction Design and continous variable In EdgeR
1
0
Entering edit mode
2.2 years ago

Hi,

I have a question about the GLM options of edgeR, I have a set-up where individuals were subjected to different condition (Continuous) and the samples are of two type. I aim to look at the interaction of those two term. However the design from the manual seem odd to me (however I am relatively new to these statistical analysis) :

(Intercept) Condition Type1 Condition:Type1

1 1 1 1 1

...

6 1 1 1 1

7 1 2 1 2

...

12 1 2 1 2

13 1 3 1 3

..

18 1 3 1 3

19 1 1 0 0

...

24 1 1 0 0

25 1 2 0 0

...

30 1 2 0 0

31 1 3 0 0

..

36 1 3 0 0

Type<- factor(c(rep("Type1",12),rep("Type2",12)))
Condition<- (c(rep(1,4),rep(2,4),rep(3,4),rep(1,4),rep(2,4),rep(3,4)))
design <- model.matrix(~ Condition*Type, data=count)
lrt <- glmQLFTest(fit, coef = 4)

The last column strikes me as odd since the type one does not have a factor indicating there condition. Can someone help me interpret this model and what that last column test ?

My initial thinking was an interaction was that the type would influence how the condition affect the transcriptomic data (in my case). However that last column make me think the last one might test the reaction for one type but not the other...

RNA-Seq R rna-seq next-gen • 616 views
ADD COMMENT
0
Entering edit mode
2.2 years ago

Type1 doesn't need a non-zero value in the last column because it serves as the reference level. In other words, the interaction effect is relative to it.

ADD COMMENT

Login before adding your answer.

Traffic: 1593 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