Suppose you have 5 experimental conditions: no treatment control, compound A treatment, compound B treatment, compound A + B treatment, compound A_covalently_linked_to_B treatment. Or simply: control, A, B, A+B, A_B. All compounds are at the same concentration, e.g., 1uM.
Your job is to compare A_B ~ A or A_B ~ B. You can do 2 separate pair-wise comparisons: A_B ~ A, and A_B ~ B, or you can do 1 comparison.
For pair-wise comparison, the contrast is simple: contrast=c(..., -1, ..., 1) with -1 corresponding to A and 1 corresponding to A_B. Likewise use a contrast for A_B ~ B.
My question is to use one comparison and what contrast you should use: -0.5 for A, -0.5 for B, and 1 for A_B? I guess it's a rule that your contrast should sum up to 0. But intuitively I feel like to use -1 for A, -1 for B, and 1 for A_B because you are subtracting the effect of A and B from A_B, all with the same weight 1 because their concentrations are the same.
Or do I have to use the pair-wise contrast?
Thanks!
