Hello everybody,
I am fairly new to microarray expression analyses, therefore I am trying to reproduce data from raw data provided from a different research experiment. In the paper it states: "[...] the mean spot intensities were normalized using quantile normalization and the two dye-swapped profiles per sample were merged by averaging the fold changes" (Koot et al., Nature 2016).
My questions are as follows:
1. Is the quantile normalization done only on the mean values of the channels' signals together (Ch1 and Ch2) , shouldn't we standardize the means with regards to the background signal as well and then preform normalization?
2. Knowing that we have a seperate microarray for each sample, how will we obtain the final data, in other words how will we generate the fold change? Will it be from the means again only?
If anyone is interested with the data and publication I am leaving a link also.
Thank you so much
https://www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-2591/
Mr Yi,
Thank you for your answer, but I am not sure what you mean by your answer, I could'nt understand the relation between your answer and my question. Is it possible for you to explain in further detail?
It sounds like the spot intensities are already in relation to the background intensity, so you would simply use their intensity quantiles and come up with a normalized (i.e., standardized) score based on these intensities. I assume there are enough quantiles to do this viably if the intensities are not on a continuous scale. See: https://en.wikipedia.org/wiki/Quantile_normalization .
For averaging the -fold changes, it's just an average of the two. If one is a six-fold change and the other is a two-fold change, the average would be four-fold.
- Badges
- Science topic