# Example of the Contrast.txt file:
sampleB
sampleC
sampleD
sampleD-(sampleB-sampleC)/2

# ----------------------------------------------------------------------------------------------
To clarify this a little, we will consider a hypothetical experiment with four groups. The
groups correspond to four different types of cells: white and smooth, white and furry, red
and smooth and red furry. We will think of white and red as being the major group, and
smooth and furry as being a sub-grouping. Suppose the RNA samples look like this:
Sample Color Type Group
1 White	Smooth	A
2 White	Smooth	A
3 White	Furry	B
4 White	Furry	B
5 Red	Smooth	C
6 Red	Smooth	C
7 Red	Furry	D
8 Red 	Furry	D
To decide which contrasts should be made between the four groups, we need to be clear
what are our scientific hypotheses. In other words, what are we seeking to show?
First, suppose that we wish to find genes that are always higher in red cells than in white
cells. Then we will need to form the four contrasts C-A, C-B, D-A and D-B, and select genes
that are significantly up for all four contrasts.
Or suppose we wish to establish that the difference between Red and White is large
compared to the differences between Furry and Smooth. An efficient way to establish this
would be to form the three contrasts B-A, D-C and (C+D)/2-(A+B)/2. We could confidently
make this assertion for genes for which the third contrast is far more significant than the
first two. Even if B-A and D-C are statistically significant, we could still look for genes for
which the fold changes for (C+D)/2-(A+B)/2 are much larger than those for B-A or D-C.

We might want to find genes that more highly expressed in Furry cells regardless of color.
Then we would test the contrasts B-A and D-C, and look for genes that are significantly up
for both contrasts.
Or we want to assert that the difference between Furry over Smooth is much the same
regardless of color. In that case you need to show that the contrast (B+D)/2-(A+C)/2 (the
average Furry effect) is significant for many genes but that (D-C)-(B-A) (the interaction) is
not.

reference: EdgeR user guide
# ----------------------------------------------------------------------------------------------