Entering edit mode

12 weeks ago

SGMS
▴
130

Hi everyone,

In R, I want to do exactly what limma package does for "makeContrasts" (I do not have expression data).

I have two factor variables say X (levels=X1,X2) and Y (levels=Y1,Y2). I ran an lme model with an interaction term between X and Y and used the "emmeans" package to get contrasts. The emmeans though return e.g:

```
contrast estimate SE df t.ratio p.value
X1 - Y1 -0.141 1.58 180 -0.090 0.9997
X1 - X2 1.949 1.03 175 1.894 0.2322
Y1 - Y2 2.239 1.01 175 2.208 0.1228
X2 - Y2 0.148 1.57 175 0.095 0.9997
```

What I need now, is the contrast of (Y1-Y2) - (X1-X2). Is this possible using emmeans or in base R? How do I create the contrasts table along with p-values etc?

Any help would be greatly appreciated.

Thanks in advance

have you looked into

`model.matrix`

?