This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Construct correct contrast

Hello,

I was reading edgeR user guide and wasn't able to understand one of the comparisons made in the section 3.3.1.

> design
   Drug.0h Drug.1h Drug.2h Placebo.0h Placebo.1h Placebo.2h
1        0       0       0          1          0          0
2        0       0       0          1          0          0
3        0       0       0          0          1          0
4        0       0       0          0          1          0
5        0       0       0          0          0          1
6        0       0       0          0          0          1
7        1       0       0          0          0          0
8        1       0       0          0          0          0
9        0       1       0          0          0          0
10       0       1       0          0          0          0
11       0       0       1          0          0          0
12       0       0       1          0          0          0

my.contrasts <- makeContrasts(
+ DrugvsPlacebo.0h = Drug.0h-Placebo.0h,
+ DrugvsPlacebo.1h = (Drug.1h-Drug.0h)-(Placebo.1h-Placebo.0h),
+ DrugvsPlacebo.2h = (Drug.2h-Drug.0h)-(Placebo.2h-Placebo.0h),
+ levels=design)

The question is:

## To find genes that have responded differently to the drug and the placebo at 2 hours - this is from edgeR user guide.
> qlf <- glmQLFTest(fit, contrast=my.contrasts[,"DrugvsPlacebo.2h"])

why do we need to subtract 0h for each treatment when comparing the drug and placebo at 2 hours?

I thought the correct contrast would be Drug.2h - Placebo.2h ...

Thanks!

contrast model-matrix design-matrix
• 914 views • link
written by jkim • jkim

To maybe add more context, Drug at 0h and Placebo at 0h should be equivalent samples, since they are effectively both "untreated".

However, due to the way samples are collected or prepared, the 0h within Drug and Placebo may have batch effects. So, by subtracting the 0h points respectively, you are effectively controlling for any differences in the "untreated" sample for each condition.

You could do Drug.2h - Placebo.2h, and that would give you differences between the two conditions at that timepoint, but it would not account for any differences already apparent at 0 h.

• 0 views • link
written by rfran010 • rfran010

Thank you for the explanation!

• 0 views • link
written by jkim • jkim

1 answer

Both the drug and placebo must be brought to the common frame of reference (0h) before they can be compared to each other. If you did Drug.2h - Placebo.2h your frame of reference would be Placebo.2h and there would be nothing left for comparison. Your temporal frame of reference must be when the whole experiment started, and you will make conclusions based on how the two treatments differ after the same amount of time has passed for each.

• 0 views • link
written by Mensur Dlakic Mensur Dlakic

Log in to answer this question.

More posts like this