I am analyzing two-color HEEBO microarrays (from Stanford) by using Limma package. I want to perform a time course analysis but I hybridized 8 time points on 4 slides. Now, I need to separate the R and G signals in order to perform the time course analysis. After the quantile normalization between arrays I am using the MA values to transform back to R and G. Is there a better approach to get the normalized R and G values?
The way is essentially correct. You might also think of a quantile-normalization of the R and G profiles directly (you will only need the MA values when the normalization of the M- or A-values is of particular interest).
I hope you did not always label the control condition in the same color. In this case it wouldn't be possible to estimate dye-effects.
If your hybridizations are pretty much different, you might to consider a paired analysis. Simply specify the "array ID" (or the filename) as an additional predictor in the linear model.
Thank you for the answers. I will try the quantile normalization of the R and G values.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools