I have a 2 factor design with one factor being 2 levels and the other being 4 levels. I am only interested in the upregulated genes specific to the interaction of level 2 of the first factor and the 4th level of the second factor. I am unsure how to program my contrast matrix after fitting limma to get the DEGs I m interested in.

I have the following design matrix:

```
mm_noreps.interactions <- model.matrix(~condition*TRAPed)
```

Both variables are factors condition has 4 levels and TRAPed has 2 levels.

```
pheatmap(mm_noreps.interactions)
```

I want to use the following contrasts: 1) TRAP positive (just consider this a drug treatment) vs TRAP negative in all conditions, c(-1,0,0,0,0,1,0,0) 2) FearRecallpositive vs all other condition's positive cells c(0,0,-1/3,0,-1/3,-1/3,0,1), and 3) FearRecallpostive vs all other conditions c(-1/7,-1/7,-1/7,-1/7,-1/7,-1/7,-1/7,1).

The main question I have is are there differentially expressed genes specific to the FearRecall:TRAPedpostive interaction. Does contrast 3 address this? The other two are more sanity checks than actual contrasts of interest.

I am not knowledgable enough to be giving out advice, but I asked a question similar to this a few months back and Gordyn Smith recommended the

`decideTests()`

function which I did not know about previously--that function might be worth checking out?How to make contrasts to find genes DE in one experimental group but NOT in others?